Diff of /Sprout/ERDB.pm

-revision 1.16, Tue Jun 28 23:51:06 2005 UTC
+revision 1.104, Mon Sep 22 20:33:30 2008 UTC
 Line 6
      use Data::Dumper;
      use XML::Simple;
      use DBQuery;
-     use DBObject;
+     use ERDBObject;
      use Stats;
      use Time::HiRes qw(gettimeofday);
+     use Digest::MD5 qw(md5_base64);
+     use CGI;
+     use WikiTools;
  =head1 Entity-Relationship Database Package
-Line 57
+Line 60
  B<start-position>, which indicates where in the contig that the sequence begins. This attribute
  is implemented as the C<start_position> field in the C<IsMadeUpOf> relation.
- The database itself is described by an XML file using the F<ERDatabase.xsd> schema. In addition to
+ The database itself is described by an XML file. In addition to all the data required to define
- all the data required to define the entities, relationships, and attributes, the schema provides
+ the entities, relationships, and attributes, the schema provides space for notes describing
- space for notes describing the data and what it means. These notes are used by L</ShowMetaData>
+ the data and what it means. These notes are used by L</ShowMetaData> to generate documentation
- to generate documentation for the database.
+ for the database.
+ Special support is provided for text searching. An entity field can be marked as <em>searchable</em>,
+ in which case it will be used to generate a text search index in which the user searches for words
+ in the field instead of a particular field value.
  Finally, every entity and relationship object has a flag indicating if it is new or old. The object
  is considered I<old> if it was loaded by the L</LoadTables> method. It is considered I<new> if it
  was inserted by the L</InsertObject> method.
- To facilitate testing, the ERDB module supports automatic generation of test data. This process
- is described in the L</GenerateEntity> and L</GenerateConnection> methods, though it is not yet
- fully implemented.
  =head2 XML Database Description
  =head3 Data Types
-Line 89
+Line 92
 -bit signed integer
+ =item counter
+-bit unsigned integer
  =item date
 -bit unsigned integer, representing a PERL date/time value
  =item text
- long string; Text fields cannot be used in indexes or sorting and do not support the
+ long string; Text fields do not support the normal syntax of filter clauses,
- normal syntax of filter clauses, but can be up to a billion character in length
+ but can be up to a billion character in length
+ =item dna
+ long string, used to store DNA and protein sequences
+ =item image
+ long string, used to store encoded image data
  =item float
-Line 108
+Line 123
  compatability with certain database packages), but the only values supported are
 and 1.
+ =item id-string
+ variable-length string, maximum 25 characters
  =item key-string
  variable-length string, maximum 40 characters
-Line 124
+Line 143
  variable-length string, maximum 255 characters
+ =item hash-string
+ variable-length string, maximum 22 characters
  =back
+ The hash-string data type has a special meaning. The actual key passed into the loader will
+ be a string, but it will be digested into a 22-character MD5 code to save space. Although the
+ MD5 algorithm is not perfect, it is extremely unlikely two strings will have the same
+ digest. Therefore, it is presumed the keys will be unique. When the database is actually
+ in use, the hashed keys will be presented rather than the original values. For this reason,
+ they should not be used for entities where the key is meaningful.
  =head3 Global Tags
  The entire database definition must be inside a B<Database> tag. The display name of
-Line 169
+Line 199
  Name of the field. The field name should contain only letters, digits, and hyphens (C<->),
  and the first character should be a letter. Most underlying databases are case-insensitive
- with the respect to field names, so a best practice is to use lower-case letters only.
+ with the respect to field names, so a best practice is to use lower-case letters only. Finally,
+ the name C<search-relevance> has special meaning for full-text searches and should not be
+ used as a field name.
  =item type
-Line 188
+Line 220
  entity, the fields without a relation attribute are said to belong to the
  I<primary relation>. This relation has the same name as the entity itself.
+ =item searchable
+ If specified, then the field is a candidate for full-text searching. A single full-text
+ index will be created for each relation with at least one searchable field in it.
+ For best results, this option should only be used for string or text fields.
+ =item special
+ This attribute allows the subclass to assign special meaning for certain fields.
+ The interpretation is up to the subclass itself. Currently, only entity fields
+ can have this attribute.
  =back
  =head3 Indexes
- An entity can have multiple alternate indexes associated with it. The fields must
+ An entity can have multiple alternate indexes associated with it. The fields in an
- be from the primary relation. The alternate indexes assist in ordering results
+ index must all be from the same relation. The alternate indexes assist in searching
- from a query. A relationship can have up to two indexes-- a I<to-index> and a
+ on fields other than the entity ID. A relationship has at least two indexes-- a I<to-index> and a
- I<from-index>. These order the results when crossing the relationship. For
+ I<from-index> that order the results when crossing the relationship. For
  example, in the relationship C<HasContig> from C<Genome> to C<Contig>, the
  from-index would order the contigs of a ganome, and the to-index would order
- the genomes of a contig. A relationship's index must specify only fields in
+ the genomes of a contig. In addition, it can have zero or more alternate
+ indexes. A relationship's index must specify only fields in
  the relationship.
- The indexes for an entity must be listed inside the B<Indexes> tag. The from-index
+ The alternate indexes for an entity or relationship must be listed inside the B<Indexes> tag.
- of a relationship is specified using the B<FromIndex> tag; the to-index is specified
+ The from-index of a relationship is specified using the B<FromIndex> tag; the to-index is
- using the B<ToIndex> tag.
+ specified using the B<ToIndex> tag.
  Each index can contain a B<Notes> tag. In addition, it will have an B<IndexFields>
  tag containing the B<IndexField> tags. These specify, in order, the fields used in
-Line 221
+Line 266
  =back
- The B<Index>, B<FromIndex>, and B<ToIndex> tags themselves have no attributes.
+ The B<FromIndex>, and B<ToIndex> tags have no attributes. The B<Index> tag can
+ have a B<Unique> attribute. If specified, the index will be generated as a unique
+ index.
  =head3 Object and Field Names
-Line 265
+Line 312
  A relationship is described by the C<Relationship> tag. Within a relationship,
  there can be a C<Notes> tag, a C<Fields> tag containing the intersection data
- fields, a C<FromIndex> tag containing the from-index, and a C<ToIndex> tag containing
+ fields, a C<FromIndex> tag containing the from-index, a C<ToIndex> tag containing
- the to-index.
+ the to-index, and an C<Indexes> tag containing the alternate indexes.
  The C<Relationship> tag has the following attributes.
-Line 299
+Line 346
  # Table of information about our datatypes. "sqlType" is the corresponding SQL datatype string.
  # "maxLen" is the maximum permissible length of the incoming string data used to populate a field
- # of the specified type. "dataGen" is PERL string that will be evaluated if no test data generation
+ # of the specified type. "avgLen" is the average byte length for estimating
-  #string is specified in the field definition.
+ # record sizes. "sort" is the key modifier for the sort command, "notes" is a type description,
- my %TypeTable = ( char =>    { sqlType => 'CHAR(1)',            maxLen => 1,            dataGen => "StringGen('A')" },
+ # and "indexMod", if non-zero, is the number of characters to use when the field is specified in an
-                   int =>     { sqlType => 'INTEGER',            maxLen => 20,           dataGen => "IntGen(0, 99999999)" },
+ # index
-                   string =>  { sqlType => 'VARCHAR(255)',       maxLen => 255,          dataGen => "StringGen(IntGen(10,250))" },
+ my %TypeTable = ( char =>    { sqlType => 'CHAR(1)',            maxLen => 1,            avgLen =>   1, sort => "",
-                   text =>    { sqlType => 'TEXT',               maxLen => 1000000000,   dataGen => "StringGen(IntGen(80,1000))" },
+                                indexMod =>   0, notes => "single ASCII character"},
-                   date =>    { sqlType => 'BIGINT',             maxLen => 80,           dataGen => "DateGen(-7, 7, IntGen(0,1400))" },
+                   int =>     { sqlType => 'INTEGER',            maxLen => 20,           avgLen =>   4, sort => "n",
-                   float =>   { sqlType => 'DOUBLE PRECISION',   maxLen => 40,           dataGen => "FloatGen(0.0, 100.0)" },
+                                indexMod =>   0, notes => "signed 32-bit integer"},
-                   boolean => { sqlType => 'SMALLINT',           maxLen => 1,            dataGen => "IntGen(0, 1)" },
+                   counter => { sqlType => 'INTEGER UNSIGNED',   maxLen => 20,           avgLen =>   4, sort => "n",
+                                indexMod =>   0, notes => "unsigned 32-bit integer"},
+                   image =>    { sqlType => 'TEXT',               maxLen => 1000000000,   avgLen => 100000, sort => "",
+                                indexMod => 255, notes => "UUencoded image, suitable for import into GD, should never be indexed"},
+                   dna =>      { sqlType => 'TEXT',               maxLen => 1000000000,   avgLen => 100000, sort => "",
+                                indexMod => 255, notes => "DNA or protein sequence, should never be indexed"},
+                   string =>  { sqlType => 'VARCHAR(255)',       maxLen => 255,          avgLen => 100, sort => "",
+                                indexMod =>   0, notes => "character string, 0 to 255 characters"},
+                   text =>    { sqlType => 'TEXT',               maxLen => 1000000000,   avgLen => 500, sort => "",
+                                indexMod => 255, notes => "character string, nearly unlimited length, only first 255 characters are indexed"},
+                   date =>    { sqlType => 'BIGINT',             maxLen => 80,           avgLen =>   8, sort => "n",
+                                indexMod =>   0, notes => "signed, 64-bit integer"},
+                   float =>   { sqlType => 'DOUBLE PRECISION',   maxLen => 40,           avgLen =>   8, sort => "g",
+                                indexMod =>   0, notes => "64-bit double precision floating-point number"},
+                   boolean => { sqlType => 'SMALLINT',           maxLen => 1,            avgLen =>   1, sort => "n",
+                                indexMod =>   0, notes => "boolean value: 0 if false, 1 if true"},
+                  'hash-string' =>
+                              { sqlType => 'VARCHAR(22)',        maxLen => 22,           avgLen =>  22, sort => "",
+                                indexMod =>   0, notes => "string stored in digested form, used for certain types of key fields"},
+                  'id-string' =>
+                              { sqlType => 'VARCHAR(25)',        maxLen => 25,           avgLen =>  25, sort => "",
+                                indexMod =>   0, notes => "character string, 0 to 25 characters"},
                   'key-string' =>
-                              { sqlType => 'VARCHAR(40)',        maxLen => 40,           dataGen => "StringGen(IntGen(10,40))" },
+                              { sqlType => 'VARCHAR(40)',        maxLen => 40,           avgLen =>  10, sort => "",
+                                indexMod =>   0, notes => "character string, 0 to 40 characters"},
                   'name-string' =>
-                              { sqlType => 'VARCHAR(80)',        maxLen => 80,           dataGen => "StringGen(IntGen(10,80))" },
+                              { sqlType => 'VARCHAR(80)',        maxLen => 80,           avgLen =>  40, sort => "",
+                                indexMod =>   0, notes => "character string, 0 to 80 characters"},
                   'medium-string' =>
-                              { sqlType => 'VARCHAR(160)',       maxLen => 160,          dataGen => "StringGen(IntGen(10,160))" },
+                              { sqlType => 'VARCHAR(160)',       maxLen => 160,          avgLen =>  40, sort => "",
+                                indexMod =>   0, notes => "character string, 0 to 160 characters"},
+                  'long-string' =>
+                              { sqlType => 'VARCHAR(500)',       maxLen => 500,          avglen => 255, sort => "",
+                                indexMod =>   0, notes => "character string, 0 to 500 characters"},
                  );
  # Table translating arities into natural language.
-Line 322
+Line 396
                     'MM' => 'many-to-many'
                   );
- # Table for interpreting string patterns.
+ # Options for XML input and output.
- my %PictureTable = ( 'A' => "abcdefghijklmnopqrstuvwxyz",
+ my %XmlOptions = (GroupTags =>  { Relationships => 'Relationship',
-                      '9' => "0123456789",
+                                   Entities => 'Entity',
-                      'X' => "abcdefghijklmnopqrstuvwxyz0123456789",
+                                   Fields => 'Field',
-                      'V' => "aeiou",
+                                   Indexes => 'Index',
-                      'K' => "bcdfghjklmnoprstvwxyz"
+                                   IndexFields => 'IndexField',
+                                   Issues => 'Issue',
+                                   Shapes => 'Shape'
+                                 },
+                   KeyAttr =>    { Relationship => 'name',
+                                   Entity => 'name',
+                                   Field => 'name',
+                                   Shape => 'name'
+                                 },
+                   SuppressEmpty => 1,
+                  );
+ my %XmlInOpts  = (
+                   ForceArray => [qw(Field Index IndexField Relationship Entity Shape)],
+                   ForceContent => 1,
+                   NormalizeSpace => 2,
+                  );
+ my %XmlOutOpts = (
+                   RootName => 'Database',
+                   XMLDecl => 1,
                     );
  =head2 Public Methods
  =head3 new
- C<< my $database = ERDB->new($dbh, $metaFileName); >>
+     my $database = ERDB->new($dbh, $metaFileName);
  Create a new ERDB object.
-Line 355
+Line 448
  sub new {
      # Get the parameters.
-     my ($class, $dbh, $metaFileName, $options) = @_;
+     my ($class, $dbh, $metaFileName, %options) = @_;
      # Load the meta-data.
      my $metaData = _LoadMetaData($metaFileName);
      # Create the object.
-Line 369
+Line 462
  =head3 ShowMetaData
- C<< $database->ShowMetaData($fileName); >>
+     $erdb->ShowMetaData($fileName);
  This method outputs a description of the database. This description can be used to help users create
  the data to be loaded into the relations.
-Line 400
+Line 493
      # Write the HTML heading stuff.
      print HTMLOUT "<html>\n<head>\n<title>$title</title>\n";
      print HTMLOUT "</head>\n<body>\n";
+     # Write the documentation.
+     print HTMLOUT $self->DisplayMetaData();
+     # Close the document.
+     print HTMLOUT "</body>\n</html>\n";
+     # Close the file.
+     close HTMLOUT;
+ }
+ =head3 DisplayMetaData
+     my $html = $erdb->DisplayMetaData();
+ Return an HTML description of the database. This description can be used to help users create
+ the data to be loaded into the relations and form queries. The output is raw includable HTML
+ without any HEAD or BODY tags.
+ =over 4
+ =item filename
+ The name of the output file.
+ =back
+ =cut
+ sub DisplayMetaData {
+     # Get the parameters.
+     my ($self) = @_;
+     # Get the metadata and the title string.
+     my $metadata = $self->{_metaData};
+     # Get the title string.
+     my $title = $metadata->{Title};
+     # Get the entity and relationship lists.
+     my $entityList = $metadata->{Entities};
+     my $relationshipList = $metadata->{Relationships};
+     # Declare the return variable.
+     my $retVal = "";
+     # Open the output file.
+     Trace("Building MetaData table of contents.") if T(4);
      # Here we do the table of contents. It starts as an unordered list of section names. Each
      # section contains an ordered list of entity or relationship subsections.
-     print HTMLOUT "<ul>\n<li><a href=\"#EntitiesSection\">Entities</a>\n<ol>\n";
+     $retVal .= "<ul>\n<li><a href=\"#EntitiesSection\">Entities</a>\n<ol>\n";
      # Loop through the Entities, displaying a list item for each.
      foreach my $key (sort keys %{$entityList}) {
          # Display this item.
-         print HTMLOUT "<li><a href=\"#$key\">$key</a></li>\n";
+         $retVal .= "<li><a href=\"#$key\">$key</a></li>\n";
      }
      # Close off the entity section and start the relationship section.
-     print HTMLOUT "</ol></li>\n<li><a href=\"#RelationshipsSection\">Relationships</a>\n<ol>\n";
+     $retVal .= "</ol></li>\n<li><a href=\"#RelationshipsSection\">Relationships</a>\n<ol>\n";
      # Loop through the Relationships.
      foreach my $key (sort keys %{$relationshipList}) {
          # Display this item.
          my $relationshipTitle = _ComputeRelationshipSentence($key, $relationshipList->{$key});
-         print HTMLOUT "<li><a href=\"#$key\">$relationshipTitle</a></li>\n";
+         $retVal .= "<li><a href=\"#$key\">$relationshipTitle</a></li>\n";
      }
      # Close off the relationship section and list the join table section.
-     print HTMLOUT "</ol></li>\n<li><a href=\"#JoinTable\">Join Table</a></li>\n";
+     $retVal .= "</ol></li>\n<li><a href=\"#JoinTable\">Join Table</a></li>\n";
      # Close off the table of contents itself.
-     print HTMLOUT "</ul>\n";
+     $retVal .=  "</ul>\n";
      # Now we start with the actual data. Denote we're starting the entity section.
-     print HTMLOUT "<a name=\"EntitiesSection\"></a><h2>Entities</h2>\n";
+     $retVal .= "<a name=\"EntitiesSection\"></a><h2>Entities</h2>\n";
      # Loop through the entities.
      for my $key (sort keys %{$entityList}) {
          Trace("Building MetaData entry for $key entity.") if T(4);
          # Create the entity header. It contains a bookmark and the entity name.
-         print HTMLOUT "<a name=\"$key\"></a><h3>$key</h3>\n";
+         $retVal .= "<a name=\"$key\"></a><h3>$key</h3>\n";
          # Get the entity data.
          my $entityData = $entityList->{$key};
          # If there's descriptive text, display it.
          if (my $notes = $entityData->{Notes}) {
-             print HTMLOUT "<p>" . _HTMLNote($notes->{content}) . "</p>\n";
+             $retVal .= "<p>" . HTMLNote($notes->{content}) . "</p>\n";
          }
-         # Now we want a list of the entity's relationships. First, we set up the relationship subsection.
+         # See if we need a list of the entity's relationships.
-         print HTMLOUT "<h4>Relationships for <b>$key</b></h4>\n<ul>\n";
+         my $relCount = keys %{$relationshipList};
+         if ($relCount > 0) {
+             # First, we set up the relationship subsection.
+             $retVal .= "<h4>Relationships for <b>$key</b></h4>\n<ul>\n";
          # Loop through the relationships.
          for my $relationship (sort keys %{$relationshipList}) {
              # Get the relationship data.
-Line 444
+Line 580
                  # Get the relationship sentence and append the arity.
                  my $relationshipDescription = _ComputeRelationshipSentence($relationship, $relationshipStructure);
                  # Display the relationship data.
-                 print HTMLOUT "<li><a href=\"#$relationship\">$relationshipDescription</a></li>\n";
+                     $retVal .= "<li><a href=\"#$relationship\">$relationshipDescription</a></li>\n";
              }
          }
          # Close off the relationship list.
-         print HTMLOUT "</ul>\n";
+             $retVal .= "</ul>\n";
+         }
          # Get the entity's relations.
          my $relationList = $entityData->{Relations};
          # Create a header for the relation subsection.
-         print HTMLOUT "<h4>Relations for <b>$key</b></h4>\n";
+         $retVal .= "<h4>Relations for <b>$key</b></h4>\n";
          # Loop through the relations, displaying them.
          for my $relation (sort keys %{$relationList}) {
              my $htmlString = _ShowRelationTable($relation, $relationList->{$relation});
-             print HTMLOUT $htmlString;
+             $retVal .= $htmlString;
          }
      }
      # Denote we're starting the relationship section.
-     print HTMLOUT "<a name=\"RelationshipsSection\"></a><h2>Relationships</h2>\n";
+     $retVal .= "<a name=\"RelationshipsSection\"></a><h2>Relationships</h2>\n";
      # Loop through the relationships.
      for my $key (sort keys %{$relationshipList}) {
          Trace("Building MetaData entry for $key relationship.") if T(4);
-Line 468
+Line 605
          my $relationshipStructure = $relationshipList->{$key};
          # Create the relationship header.
          my $headerText = _ComputeRelationshipHeading($key, $relationshipStructure);
-         print HTMLOUT "<h3><a name=\"$key\"></a>$headerText</h3>\n";
+         $retVal .= "<h3><a name=\"$key\"></a>$headerText</h3>\n";
          # Get the entity names.
          my $fromEntity = $relationshipStructure->{from};
          my $toEntity = $relationshipStructure->{to};
-Line 478
+Line 615
          # since both sentences will say the same thing.
          my $arity = $relationshipStructure->{arity};
          if ($arity eq "11") {
-             print HTMLOUT "<p>Each <b>$fromEntity</b> relates to at most one <b>$toEntity</b>.\n";
+             $retVal .= "<p>Each <b>$fromEntity</b> relates to at most one <b>$toEntity</b>.\n";
          } else {
-             print HTMLOUT "<p>Each <b>$fromEntity</b> relates to multiple <b>$toEntity</b>s.\n";
+             $retVal .= "<p>Each <b>$fromEntity</b> relates to multiple <b>$toEntity</b>s.\n";
              if ($arity eq "MM" && $fromEntity ne $toEntity) {
-                 print HTMLOUT "Each <b>$toEntity</b> relates to multiple <b>$fromEntity</b>s.\n";
+                 $retVal .= "Each <b>$toEntity</b> relates to multiple <b>$fromEntity</b>s.\n";
              }
          }
-         print HTMLOUT "</p>\n";
+         $retVal .= "</p>\n";
          # If there are notes on this relationship, display them.
          if (my $notes = $relationshipStructure->{Notes}) {
-             print HTMLOUT "<p>" . _HTMLNote($notes->{content}) . "</p>\n";
+             $retVal .= "<p>" . HTMLNote($notes->{content}) . "</p>\n";
          }
          # Generate the relationship's relation table.
          my $htmlString = _ShowRelationTable($key, $relationshipStructure->{Relations}->{$key});
-         print HTMLOUT $htmlString;
+         $retVal .= $htmlString;
      }
      Trace("Building MetaData join table.") if T(4);
      # Denote we're starting the join table.
-     print HTMLOUT "<a name=\"JoinTable\"></a><h3>Join Table</h3>\n";
+     $retVal .= "<a name=\"JoinTable\"></a><h3>Join Table</h3>\n";
      # Create a table header.
-     print HTMLOUT _OpenTable("Join Table", "Source", "Target", "Join Condition");
+     $retVal .= _OpenTable("Join Table", "Source", "Target", "Join Condition");
      # Loop through the joins.
      my $joinTable = $metadata->{Joins};
      my @joinKeys = keys %{$joinTable};
-Line 506
+Line 643
          # Separate out the source, the target, and the join clause.
          $joinKey =~ m!^([^/]+)/(.+)$!;
          my ($sourceRelation, $targetRelation) = ($1, $2);
-         Trace("Join with key $joinKey is from $sourceRelation to $targetRelation.") if T(4);
+         Trace("Join with key $joinKey is from $sourceRelation to $targetRelation.") if T(Joins => 4);
          my $source = $self->ComputeObjectSentence($sourceRelation);
          my $target = $self->ComputeObjectSentence($targetRelation);
          my $clause = $joinTable->{$joinKey};
          # Display them in a table row.
-         print HTMLOUT "<tr><td>$source</td><td>$target</td><td>$clause</td></tr>\n";
+         $retVal .= "<tr><td>$source</td><td>$target</td><td>$clause</td></tr>\n";
      }
      # Close the table.
-     print HTMLOUT _CloseTable();
+     $retVal .= _CloseTable();
-     # Close the document.
+     Trace("Built MetaData HTML.") if T(3);
-     print HTMLOUT "</body>\n</html>\n";
+     # Return the HTML.
-     # Close the file.
+     return $retVal;
-     close HTMLOUT;
-     Trace("Built MetaData web page.") if T(3);
  }
  =head3 DumpMetaData
- C<< $database->DumpMetaData(); >>
+     $erdb->DumpMetaData();
  Return a dump of the metadata structure.
-Line 537
+Line 672
      return Data::Dumper::Dumper($self->{_metaData});
  }
+ =head3 GenerateWikiData
+     my @wikiLines = $erdb->GenerateWikiData();
+ Build a description of the database for the wiki. The database will be
+ organized into a single page, with sections for each entity and relationship.
+ The return value is a list of text lines.
+ =cut
+ sub GenerateWikiData {
+     # Get the parameters.
+     my ($self) = @_;
+     # We'll build the wiki text in here.
+     my @retVal = ();
+     # Get the metadata object.
+     my $metadata = $self->{_metaData};
+     # Get the title string. This will become the page name.
+     my $title = $metadata->{Title}->{content};
+     # Get the entity and relationship lists.
+     my $entityList = $metadata->{Entities};
+     my $relationshipList = $metadata->{Relationships};
+     my $shapeList = $metadata->{Shapes};
+     # Start with the introductory text.
+     push @retVal, WikiTools::Heading(2, "Introduction");
+     if (my $notes = $metadata->{Notes}) {
+         push @retVal, WikiNote($notes->{content});
+     }
+     # Generate issue list.
+     if (my $issues = $metadata->{Issues}) {
+         push @retVal, WikiTools::Heading(3, 'Issues');
+         push @retVal, WikiTools::List(map { $_->{content} } @{$issues});
+     }
+     # Start the entity section.
+     push @retVal, WikiTools::Heading(2, "Entities");
+     # Loop through the entities. Note that unlike the situation with HTML, we
+     # don't need to generate the table of contents manually, just the data
+     # itself.
+     for my $key (sort keys %$entityList) {
+         # Create a header for this entity.
+         push @retVal, "", WikiTools::Heading(3, $key);
+         # Get the entity data.
+         my $entityData = $entityList->{$key};
+         # Plant the notes here, if there are any.
+         push @retVal, _ObjectNotes($entityData);
+         # Now we list the entity's relationships (if any). First, we build a list
+         # of the relationships relevant to this entity.
+         my @rels = ();
+         for my $rel (sort keys %$relationshipList) {
+             my $relStructure = $relationshipList->{$rel};
+             if ($relStructure->{from} eq $key || $relStructure->{to} eq $key) {
+                 # Get the relationship sentence.
+                 my $relSentence = _ComputeRelationshipSentence($rel, $relStructure);
+                 # Linkify it.
+                 my $linkedRel = WikiTools::LinkMarkup("#$rel", $rel);
+                 $relSentence =~ s/$rel/$linkedRel/;
+                 push @rels, $relSentence;
+             }
+         }
+         # Add the relationships as a Wiki list.
+         push @retVal, WikiTools::List(@rels);
+         # Get the entity's relations.
+         my $relationList = $entityData->{Relations};
+         # Loop through the relations, displaying them.
+         for my $relation (sort keys %{$relationList}) {
+             my $wikiString = _WikiRelationTable($relation, $relationList->{$relation});
+             push @retVal, $wikiString;
+         }
+     }
+     # Now the entities are documented. Next we do the relationships.
+     push @retVal, WikiTools::Heading(2, "Relationships");
+     for my $key (sort keys %$relationshipList) {
+         my $relationshipData = $relationshipList->{$key};
+         # Create the relationship heading.
+         push @retVal, WikiTools::Heading(3, $key);
+         # Describe the relationship arity. Note there's a bit of trickiness involving recursive
+         # many-to-many relationships. In a normal many-to-many we use two sentences to describe
+         # the arity (one for each direction). This is a bad idea for a recursive relationship,
+         # since both sentences will say the same thing.
+         my $arity = $relationshipData->{arity};
+         my $fromEntity = $relationshipData->{from};
+         my $toEntity = $relationshipData->{to};
+         my @listElements = ();
+         my $boldCode = WikiTools::BoldCode();
+         if ($arity eq "11") {
+             push @listElements, "Each $boldCode$fromEntity$boldCode relates to at most one $boldCode$toEntity$boldCode.";
+         } else {
+             push @listElements, "Each $boldCode$fromEntity$boldCode relates to multiple $boldCode${toEntity}s$boldCode.";
+             if ($arity eq "MM" && $fromEntity ne $toEntity) {
+                 push @listElements, "Each $boldCode$toEntity$boldCode relates to multiple $boldCode${fromEntity}s$boldCode.";
+             }
+         }
+         push @retVal, WikiTools::List(@listElements);
+         # Plant the notes here, if there are any.
+         push @retVal, _ObjectNotes($relationshipData);
+         # Finally, the relationship table.
+         my $wikiString = _WikiRelationTable($key, $relationshipData->{Relations}->{$key});
+         push @retVal, $wikiString;
+     }
+     # Now loop through the miscellaneous shapes.
+     if ($shapeList) {
+         push @retVal, WikiTools::Heading(2, "Miscellaneous");
+         for my $shape (sort keys %$shapeList) {
+             push @retVal, WikiTools::Heading(3, $shape);
+             my $shapeData = $shapeList->{$shape};
+             push @retVal, _ObjectNotes($shapeData);
+         }
+     }
+     # All done. Return the lines.
+     return @retVal;
+ }
+ =head3 CreatePPO
+     ERDB::CreatePPO($erdbXMLFile, $ppoXMLFile);
+ Create a PPO XML file from an ERDB data definition XML file. At the
+ current time, the PPO XML file can be used to create a database with
+ similar functionality. Eventually, the PPO will be able to use the
+ created XML to access the live ERDB database.
+ =over 4
+ =item erdbXMLFile
+ Name of the XML data definition file for the ERDB database. This
+ file must exist.
+ =item ppoXMLFile
+ Output file for the PPO XML definition. If this file exists, it
+ will be overwritten.
+ =back
+ =cut
+ sub CreatePPO {
+     # Get the parameters.
+     my ($erdbXMLFile, $ppoXMLFile) = @_;
+     # First, we want to slurp in the ERDB XML file in its raw form.
+     my $xml = ReadMetaXML($erdbXMLFile);
+     # Create a variable to hold all of the objects in the PPO project.
+     my @objects = ();
+     # Get the relationship hash.
+     my $relationships = $xml->{Relationships};
+     # Loop through the entities.
+     my $entities = $xml->{Entities};
+     for my $entityName (keys %{$entities}) {
+         # Get the entity's data structures.
+         my $entityObject = $entities->{$entityName};
+         # We put the object's fields in here, according to their type.
+         my (@object_refs, @scalars, @indexes, @arrays);
+         # Create the ID field for the entity. We get the key type from the
+         # entity object and compute the corresponding SQL type.
+         my $type = $TypeTable{$entityObject->{keyType}}->{sqlType};
+         push @scalars, { label => 'id', type => $type };
+         # Loop through the entity fields.
+         for my $fieldName ( keys %{$entityObject->{Fields}} ) {
+             # Get the field object.
+             my $fieldObject = $entityObject->{Fields}->{$fieldName};
+             # Convert it to a scalar tag.
+             my $scalar = _CreatePPOField($fieldName, $fieldObject);
+             # If we have a relation, this field is stored in an array.
+             # otherwise, it is a scalar. The array tag has scalars
+             # stored as an XML array. In ERDB, there is only ever one,
+             # but PPO can have more.
+             my $relation = $fieldObject->{relation};
+             if ($relation) {
+                 push @arrays, { scalar => [$scalar] };
+             } else {
+                 push @scalars, $scalar;
+             }
+         }
+         # Loop through the relationships. If this entity is the to-entity
+         # on a relationship of 1M arity, then it is implemented as a PPO
+         # object reference.
+         for my $relationshipName (keys %{$relationships}) {
+             # Get the relationship data.
+             my $relationshipData = $relationships->{$relationshipName};
+             # If we have a from for this entity and an arity of 1M, we
+             # have an object reference.
+             if ($relationshipData->{to} eq $entityName &&
+                 $relationshipData->{arity} eq '1M') {
+                 # Build the object reference tag.
+                 push @object_refs, { label => $relationshipName,
+                                      type => $relationshipData->{from} };
+             }
+         }
+         # Create the indexes.
+         my $indexList = $entityObject->{Indexes};
+         push @indexes, map { _CreatePPOIndex($_) } @{$indexList};
+         # Build the object XML tree.
+         my $object = { label => $entityName,
+                        object_ref => \@object_refs,
+                        scalar => \@scalars,
+                        index => \@indexes,
+                        array => \@arrays
+                       };
+         # Push the object onto the objects list.
+         push @objects, $object;
+     }
+     # Loop through the relationships, searching for MMs. The 1Ms were
+     # already handled by the entity search above.
+     for my $relationshipName (keys %{$relationships}) {
+         # Get this relationship's object.
+         my $relationshipObject = $relationships->{$relationshipName};
+         # Only proceed if it's many-to-many.
+         if ($relationshipObject->{arity} eq 'MM') {
+             # Create the tag lists for the relationship object.
+             my (@object_refs, @scalars, @indexes);
+             # The relationship will be created as an object with object
+             # references for its links to the participating entities.
+             my %links = ( from_link => $relationshipObject->{from},
+                           to_link => $relationshipObject->{to} );
+             for my $link (keys %links) {
+                 # Create an object_ref tag for this piece of the
+                 # relationship (from or to).
+                 my $object_ref = { label => $link,
+                                    type => $links{$link} };
+                 push @object_refs, $object_ref;
+             }
+             # Loop through the intersection data fields, creating scalar tags.
+             # There are no fancy array tags in a relationship.
+             for my $fieldName (keys %{$relationshipObject->{Fields}}) {
+                 my $fieldObject = $relationshipObject->{Fields}->{$fieldName};
+                 push @scalars, _CreatePPOField($fieldName, $fieldObject);
+             }
+             # Finally, the indexes: currently we cannot support the to-index and
+             # from-index in PPO, so we just process the alternate indexes.
+             my $indexList = $relationshipObject->{Indexes};
+             push @indexes, map { _CreatePPOIndex($_) } @{$indexList};
+             # Wrap up all the stuff about this relationship.
+             my $object = { label => $relationshipName,
+                            scalar => \@scalars,
+                            object_ref => \@object_refs,
+                            index => \@indexes
+                          };
+             # Push it into the object list.
+             push @objects, $object;
+         }
+     }
+     # Compute a title.
+     my $title;
+     if ($erdbXMLFile =~ /(\/|^)([^\/]+)DBD\.xml/) {
+         # Here we have a standard file name we can use for a title.
+         $title = $2;
+     } else {
+         # Here the file name is non-standard, so we carve up the
+         # database title.
+         $title = $xml->{Title}->{content};
+         $title =~ s/\s\.,//g;
+     }
+     # Wrap up the XML as a project.
+     my $ppoXML = { project => { label => $title,
+                                 object => \@objects }};
+     # Write out the results.
+     my $ppoString = XML::Simple::XMLout($ppoXML,
+                                         AttrIndent => 1,
+                                         KeepRoot => 1);
+     Tracer::PutFile($ppoXMLFile, [ $ppoString ]);
+ }
+ =head3 FindIndexForEntity
+     my $indexFound = ERDB::FindIndexForEntity($xml, $entityName, $attributeName);
+ This method locates the entry in an entity's index list that begins with the
+ specified attribute name. If the entity has no index list, one will be
+ created. This method works on raw XML, not a live ERDB object.
+ =over 4
+ =item xml
+ The raw XML structure defining the database.
+ =item entityName
+ The name of the relevant entity.
+ =item attributeName
+ The name of the attribute relevant to the search.
+ =item RETURN
+ The numerical index in the index list of the index entry for the specified entity and
+ attribute, or C<undef> if no such index exists.
+ =back
+ =cut
+ sub FindIndexForEntity {
+     # Get the parameters.
+     my ($xml, $entityName, $attributeName) = @_;
+     # Declare the return variable.
+     my $retVal;
+     # Get the named entity.
+     my $entityData = $xml->{Entities}->{$entityName};
+     if (! $entityData) {
+         Confess("Entity $entityName not found in DBD structure.");
+     } else {
+         # Insure it has an index list.
+         if (! exists $entityData->{Indexes}) {
+             $entityData->{Indexes} = [];
+         } else {
+             # Search for the desired index.
+             my $indexList = $entityData->{Indexes};
+             my $n = scalar @{$indexList};
+             Trace("Searching $n indexes in index list for $entityName.") if T(2);
+             # We use an indexed FOR here because we're returning an
+             # index number instead of an object. We do THAT so we can
+             # delete the index from the list if needed.
+             for (my $i = 0; $i < $n && !defined($retVal); $i++) {
+                 my $index = $indexList->[$i];
+                 my $fields = $index->{IndexFields};
+                 # Technically this IF should be safe (that is, we are guaranteed
+                 # the existence of a "$fields->[0]"), because when we load the XML
+                 # we have SuppressEmpty specified.
+                 if ($fields->[0]->{name} eq $attributeName) {
+                     $retVal = $i;
+                 }
+             }
+         }
+     }
+     Trace("Index for $attributeName of $entityName found at position $retVal.") if defined($retVal) && T(3);
+     Trace("Index for $attributeName not found in $entityName.") if !defined($retVal) && T(3);
+     # Return the result.
+     return $retVal;
+ }
  =head3 CreateTables
- C<< $datanase->CreateTables(); >>
+     $erdb->CreateTables();
  This method creates the tables for the database from the metadata structure loaded by the
  constructor. It is expected this function will only be used on rare occasions, when the
-Line 551
+Line 1020
  sub CreateTables {
      # Get the parameters.
      my ($self) = @_;
-     my $metadata = $self->{_metaData};
+     # Get the relation names.
-     my $dbh = $self->{_dbh};
+     my @relNames = $self->GetTableNames();
-     # Loop through the entities.
+     # Loop through the relations.
-     my $entityHash = $metadata->{Entities};
+     for my $relationName (@relNames) {
-     for my $entityName (keys %{$entityHash}) {
-         my $entityData = $entityHash->{$entityName};
-         # Tell the user what we're doing.
-         Trace("Creating relations for entity $entityName.") if T(1);
-         # Loop through the entity's relations.
-         for my $relationName (keys %{$entityData->{Relations}}) {
              # Create a table for this relation.
-             $self->CreateTable($relationName);
+         $self->CreateTable($relationName, 1);
-             Trace("Relation $relationName created.") if T(1);
+         Trace("Relation $relationName created.") if T(2);
-         }
-     }
-     # Loop through the relationships.
-     my $relationshipTable = $metadata->{Relationships};
-     for my $relationshipName (keys %{$metadata->{Relationships}}) {
-         # Create a table for this relationship.
-         Trace("Creating relationship $relationshipName.") if T(1);
-         $self->CreateTable($relationshipName);
      }
  }
  =head3 CreateTable
- C<< $database->CreateTable($tableName, $indexFlag); >>
+     $erdb->CreateTable($tableName, $indexFlag, $estimatedRows);
  Create the table for a relation and optionally create its indexes.
-Line 587
+Line 1042
  Name of the relation (which will also be the table name).
- =item $indexFlag
+ =item indexFlag
  TRUE if the indexes for the relation should be created, else FALSE. If FALSE,
  L</CreateIndexes> must be called later to bring the indexes into existence.
+ =item estimatedRows (optional)
+ If specified, the estimated maximum number of rows for the relation. This
+ information allows the creation of tables using storage engines that are
+ faster but require size estimates, such as MyISAM.
  =back
  =cut
  sub CreateTable {
      # Get the parameters.
-     my ($self, $relationName, $indexFlag) = @_;
+     my ($self, $relationName, $indexFlag, $estimatedRows) = @_;
      # Get the database handle.
      my $dbh = $self->{_dbh};
      # Get the relation data and determine whether or not the relation is primary.
-Line 613
+Line 1074
          # Push the result into the field list.
          push @fieldList, $fieldString;
      }
-     # If this is a root table, add the "new_record" flag. It defaults to 0, so
-     if ($rootFlag) {
-         push @fieldList, "new_record $TypeTable{boolean}->{sqlType} NOT NULL DEFAULT 0";
-     }
      # Convert the field list into a comma-delimited string.
      my $fieldThing = join(', ', @fieldList);
      # Insure the table is not already there.
      $dbh->drop_table(tbl => $relationName);
      Trace("Table $relationName dropped.") if T(2);
+     # If there are estimated rows, create an estimate so we can take advantage of
+     # faster DB technologies.
+     my $estimation = undef;
+     if ($estimatedRows) {
+         $estimation = [$self->EstimateRowSize($relationName), $estimatedRows];
+         Trace("$estimation->[1] rows of $estimation->[0] bytes each.") if T(3);
+     }
      # Create the table.
      Trace("Creating table $relationName: $fieldThing") if T(2);
-     $dbh->create_table(tbl => $relationName, flds => $fieldThing);
+     $dbh->create_table(tbl => $relationName, flds => $fieldThing, estimates => $estimation);
      Trace("Relation $relationName created in database.") if T(2);
-     # If we want to build the indexes, we do it here.
+     # If we want to build the indexes, we do it here. Note that the full-text search
+     # index will not be built until the table has been loaded.
      if ($indexFlag) {
          $self->CreateIndex($relationName);
      }
  }
+ =head3 VerifyFields
+     my $count = $erdb->VerifyFields($relName, \@fieldList);
+ Run through the list of proposed field values, insuring that all the character fields are
+ below the maximum length. If any fields are too long, they will be truncated in place.
+ =over 4
+ =item relName
+ Name of the relation for which the specified fields are destined.
+ =item fieldList
+ Reference to a list, in order, of the fields to be put into the relation.
+ =item RETURN
+ Returns the number of fields truncated.
+ =back
+ =cut
+ sub VerifyFields {
+     # Get the parameters.
+     my ($self, $relName, $fieldList) = @_;
+     # Initialize the return value.
+     my $retVal = 0;
+     # Get the relation definition.
+     my $relData = $self->_FindRelation($relName);
+     # Get the list of field descriptors.
+     my $fieldTypes = $relData->{Fields};
+     my $fieldCount = scalar @{$fieldTypes};
+     # Loop through the two lists.
+     for (my $i = 0; $i < $fieldCount; $i++) {
+         # Get the type of the current field.
+         my $fieldType = $fieldTypes->[$i]->{type};
+         # If it's a character field, verify the length.
+         if ($fieldType =~ /string/) {
+             my $maxLen = $TypeTable{$fieldType}->{maxLen};
+             my $oldString = $fieldList->[$i];
+             if (length($oldString) > $maxLen) {
+                 # Here it's too big, so we truncate it.
+                 Trace("Truncating field $i ($fieldTypes->[$i]->{name}) in relation $relName to $maxLen characters from \"$oldString\".") if T(1);
+                 $fieldList->[$i] = substr $oldString, 0, $maxLen;
+                 $retVal++;
+             }
+         }
+     }
+     # Return the truncation count.
+     return $retVal;
+ }
+ =head3 DigestFields
+     $erdb->DigestFields($relName, $fieldList);
+ Digest the strings in the field list that correspond to data type C<hash-string> in the
+ specified relation.
+ =over 4
+ =item relName
+ Name of the relation to which the fields belong.
+ =item fieldList
+ List of field contents to be loaded into the relation.
+ =back
+ =cut
+ #: Return Type ;
+ sub DigestFields {
+     # Get the parameters.
+     my ($self, $relName, $fieldList) = @_;
+     # Get the relation definition.
+     my $relData = $self->_FindRelation($relName);
+     # Get the list of field descriptors.
+     my $fieldTypes = $relData->{Fields};
+     my $fieldCount = scalar @{$fieldTypes};
+     # Loop through the two lists.
+     for (my $i = 0; $i < $fieldCount; $i++) {
+         # Get the type of the current field.
+         my $fieldType = $fieldTypes->[$i]->{type};
+         # If it's a hash string, digest it in place.
+         if ($fieldType eq 'hash-string') {
+             $fieldList->[$i] = $self->DigestKey($fieldList->[$i]);
+         }
+     }
+ }
+ =head3 DigestKey
+     my $digested = $erdb->DigestKey($keyValue);
+ Return the digested value of a symbolic key. The digested value can then be plugged into a
+ key-based search into a table with key-type hash-string.
+ Currently the digesting process is independent of the database structure, but that may not
+ always be the case, so this is an instance method instead of a static method.
+ =over 4
+ =item keyValue
+ Key value to digest.
+ =item RETURN
+ Digested value of the key.
+ =back
+ =cut
+ sub DigestKey {
+     # Get the parameters.
+     my ($self, $keyValue) = @_;
+     # Compute the digest.
+     my $retVal = md5_base64($keyValue);
+     # Return the result.
+     return $retVal;
+ }
  =head3 CreateIndex
- C<< $database->CreateIndex($relationName); >>
+     $erdb->CreateIndex($relationName);
  Create the indexes for a relation. If a table is being loaded from a large source file (as
  is the case in L</LoadTable>), it is sometimes best to create the indexes after the load.
-Line 655
+Line 1248
      for my $indexName (keys %{$indexHash}) {
          my $indexData = $indexHash->{$indexName};
          # Get the index's field list.
-         my @fieldList = _FixNames(@{$indexData->{IndexFields}});
+         my @rawFields = @{$indexData->{IndexFields}};
+         # Get a hash of the relation's field types.
+         my %types = map { $_->{name} => $_->{type} } @{$relationData->{Fields}};
+         # We need to check for text fields so we can append a length limitation for them. To do
+         # that, we need the relation's field list.
+         my $relFields = $relationData->{Fields};
+         for (my $i = 0; $i <= $#rawFields; $i++) {
+             # Get the field type.
+             my $field = $rawFields[$i];
+             my $type = $types{$field};
+             # Ask if it requires using prefix notation for the index.
+             my $mod = $TypeTable{$type}->{indexMod};
+             Trace("Field $field ($i) in $relationName has type $type and indexMod $mod.") if T(3);
+             if ($mod) {
+                 # Append the prefix length to the field name,
+                 $rawFields[$i] .= "($mod)";
+             }
+         }
+         my @fieldList = _FixNames(@rawFields);
          my $flds = join(', ', @fieldList);
          # Get the index's uniqueness flag.
-         my $unique = (exists $indexData->{Unique} ? $indexData->{Unique} : 'false');
+         my $unique = (exists $indexData->{Unique} ? 'unique' : undef);
          # Create the index.
-         $dbh->create_index(idx => $indexName, tbl => $relationName, flds => $flds, unique => $unique);
+         my $rv = $dbh->create_index(idx => $indexName, tbl => $relationName,
+                                     flds => $flds, kind => $unique);
+         if ($rv) {
          Trace("Index created: $indexName for $relationName ($flds)") if T(1);
+         } else {
+             Confess("Error creating index $indexName for $relationName using ($flds): " . $dbh->error_message());
+         }
      }
  }
- =head3 LoadTables
+ =head3 GetSecondaryFields
- C<< my $stats = $database->LoadTables($directoryName, $rebuild); >>
- This method will load the database tables from a directory. The tables must already have been created
+     my %fieldTuples = $erdb->GetSecondaryFields($entityName);
- in the database. (This can be done by calling L</CreateTables>.) The caller passes in a directory name;
- all of the relations to be loaded must have a file in the directory with the same name as the relation
- (optionally with a suffix of C<.dtx>). Each file must be a tab-delimited table of field values. Each
- line of the file will be loaded as a row of the target relation table. The field values should be in
- the same order as the fields in the relation tables generated by L</ShowMetaData>. The old data is
- erased before the new data is loaded in.
- A certain amount of translation automatically takes place. Ctrl-M characters are deleted, and
+ This method will return a list of the name and type of each of the secondary
- tab and new-line characters inside a field are escaped as C<\t> and C<\n>, respectively. Dates must
+ fields for a specified entity. Secondary fields are stored in two-column tables
- be entered as a Unix timestamp, that is, as an integer number of seconds since the base epoch.
+ in addition to the primary entity table. This enables the field to have no value
+ or to have multiple values.
  =over 4
- =item directoryName
+ =item entityName
- Name of the directory containing the relation files to be loaded.
+ Name of the entity whose secondary fields are desired.
- =item rebuild
+ =item RETURN
+ Returns a hash mapping the field names to their field types.
+ =back
+ =cut
+ sub GetSecondaryFields {
+     # Get the parameters.
+     my ($self, $entityName) = @_;
+     # Declare the return variable.
+     my %retVal = ();
+     # Look for the entity.
+     my $table = $self->GetFieldTable($entityName);
+     # Loop through the fields, pulling out the secondaries.
+     for my $field (sort keys %{$table}) {
+         if ($table->{$field}->{relation} ne $entityName) {
+             # Here we have a secondary field.
+             $retVal{$field} = $table->{$field}->{type};
+         }
+     }
+     # Return the result.
+     return %retVal;
+ }
+ =head3 GetFieldRelationName
+     my $name = $erdb->GetFieldRelationName($objectName, $fieldName);
+ Return the name of the relation containing a specified field.
+ =over 4
+ =item objectName
+ Name of the entity or relationship containing the field.
+ =item fieldName
+ Name of the relevant field in that entity or relationship.
+ =item RETURN
+ Returns the name of the database relation containing the field, or C<undef> if
+ the field does not exist.
+ =back
+ =cut
+ sub GetFieldRelationName {
+     # Get the parameters.
+     my ($self, $objectName, $fieldName) = @_;
+     # Declare the return variable.
+     my $retVal;
+     # Get the object field table.
+     my $table = $self->GetFieldTable($objectName);
+     # Only proceed if the field exists.
+     if (exists $table->{$fieldName}) {
+         # Determine the name of the relation that contains this field.
+         $retVal = $table->{$fieldName}->{relation};
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 DeleteValue
+     my $numDeleted = $erdb->DeleteValue($entityName, $id, $fieldName, $fieldValue);
+ Delete secondary field values from the database. This method can be used to delete all
+ values of a specified field for a particular entity instance, or only a single value.
+ Secondary fields are stored in two-column relations separate from an entity's primary
+ table, and as a result a secondary field can legitimately have no value or multiple
+ values. Therefore, it makes sense to talk about deleting secondary fields where it
+ would not make sense for primary fields.
+ =over 4
+ =item entityName
+ Name of the entity from which the fields are to be deleted.
+ =item id
+ ID of the entity instance to be processed. If the instance is not found, this
+ method will have no effect. If C<undef> is specified, all values for all of
+ the entity instances will be deleted.
+ =item fieldName
+ Name of the field whose values are to be deleted.
+ =item fieldValue (optional)
+ Value to be deleted. If not specified, then all values of the specified field
+ will be deleted for the entity instance. If specified, then only the values which
+ match this parameter will be deleted.
+ =item RETURN
+ Returns the number of rows deleted.
+ =back
+ =cut
+ sub DeleteValue {
+     # Get the parameters.
+     my ($self, $entityName, $id, $fieldName, $fieldValue) = @_;
+     # Declare the return value.
+     my $retVal = 0;
+     # We need to set up an SQL command to do the deletion. First, we
+     # find the name of the field's relation.
+     my $table = $self->GetFieldTable($entityName);
+     my $field = $table->{$fieldName};
+     my $relation = $field->{relation};
+     # Make sure this is a secondary field.
+     if ($relation eq $entityName) {
+         Confess("Cannot delete values of $fieldName for $entityName.");
+     } else {
+         # Set up the SQL command to delete all values.
+         my $sql = "DELETE FROM $relation";
+         # Build the filter.
+         my @filters = ();
+         my @parms = ();
+         # Check for a filter by ID.
+         if (defined $id) {
+             push @filters, "id = ?";
+             push @parms, $id;
+         }
+         # Check for a filter by value.
+         if (defined $fieldValue) {
+             push @filters, "$fieldName = ?";
+             push @parms, $fieldValue;
+         }
+         # Append the filters to the command.
+         if (@filters) {
+             $sql .= " WHERE " . join(" AND ", @filters);
+         }
+         # Execute the command.
+         my $dbh = $self->{_dbh};
+         $retVal = $dbh->SQL($sql, 0, @parms);
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 LoadTables
+     my $stats = $erdb->LoadTables($directoryName, $rebuild);
+ This method will load the database tables from a directory. The tables must already have been created
+ in the database. (This can be done by calling L</CreateTables>.) The caller passes in a directory name;
+ all of the relations to be loaded must have a file in the directory with the same name as the relation
+ (optionally with a suffix of C<.dtx>). Each file must be a tab-delimited table of field values. Each
+ line of the file will be loaded as a row of the target relation table. The field values should be in
+ the same order as the fields in the relation tables generated by L</ShowMetaData>. The old data is
+ erased before the new data is loaded in.
+ A certain amount of translation automatically takes place. Ctrl-M characters are deleted, and
+ tab and new-line characters inside a field are escaped as C<\t> and C<\n>, respectively. Dates must
+ be entered as a Unix timestamp, that is, as an integer number of seconds since the base epoch.
+ =over 4
+ =item directoryName
+ Name of the directory containing the relation files to be loaded.
+ =item rebuild
  TRUE if the tables should be dropped and rebuilt, else FALSE. This is, unfortunately, the
  only way to erase existing data in the tables, since the TRUNCATE command is not supported
-Line 710
+Line 1491
      $directoryName =~ s!/\\$!!;
      # Declare the return variable.
      my $retVal = Stats->new();
-     # Get the metadata structure.
+     # Get the relation names.
-     my $metaData = $self->{_metaData};
+     my @relNames = $self->GetTableNames();
-     # Loop through the entities.
+     for my $relationName (@relNames) {
-     for my $entity (values %{$metaData->{Entities}}) {
-         # Loop through the entity's relations.
-         for my $relationName (keys %{$entity->{Relations}}) {
              # Try to load this relation.
              my $result = $self->_LoadRelation($directoryName, $relationName, $rebuild);
              # Accumulate the statistics.
              $retVal->Accumulate($result);
          }
-     }
-     # Loop through the relationships.
-     for my $relationshipName (keys %{$metaData->{Relationships}}) {
-         # Try to load this relationship's relation.
-         my $result = $self->_LoadRelation($directoryName, $relationshipName, $rebuild);
-         # Accumulate the statistics.
-         $retVal->Accumulate($result);
-     }
      # Add the duration of the load to the statistical object.
      $retVal->Add('duration', gettimeofday - $startTime);
      # Return the accumulated statistics.
      return $retVal;
  }
  =head3 GetTableNames
- C<< my @names = $database->GetTableNames; >>
+     my @names = $erdb->GetTableNames;
  Return a list of the relations required to implement this database.
-Line 754
+Line 1525
  =head3 GetEntityTypes
- C<< my @names = $database->GetEntityTypes; >>
+     my @names = $erdb->GetEntityTypes;
  Return a list of the entity type names.
-Line 769
+Line 1540
      return sort keys %{$entityList};
  }
+ =head3 GetConnectingRelationships
+     my @list = $erdb->GetConnectingRelationships($entityName);
+ Return a list of the relationships connected to the specified entity.
+ =over 4
+ =item entityName
+ Entity whose connected relationships are desired.
+ =item RETURN
+ Returns a list of the relationships that originate from the entity.
+ If the entity is on the from end, it will return the relationship
+ name. If the entity is on the to end it will return the converse of
+ the relationship name.
+ =back
+ =cut
+ sub GetConnectingRelationships {
+     # Get the parameters.
+     my ($self, $entityName) = @_;
+     # Declare the return variable.
+     my @retVal;
+     # Get the relationship list.
+     my $relationships = $self->{_metaData}->{Relationships};
+     # Find the entity.
+     my $entity = $self->{_metaData}->{Entities}->{$entityName};
+     # Only proceed if the entity exists.
+     if (! defined $entity) {
+         Trace("Entity $entityName not found.") if T(3);
+     } else {
+         # Loop through the relationships.
+         my @rels = keys %$relationships;
+         Trace(scalar(@rels) . " relationships found in connection search.") if T(3);
+         for my $relationshipName (@rels) {
+             my $relationship = $relationships->{$relationshipName};
+             if ($relationship->{from} eq $entityName) {
+                 # Here we have a forward relationship.
+                 push @retVal, $relationshipName;
+             } elsif ($relationship->{to} eq $entityName) {
+                 # Here we have a backward relationship. In this case, the
+                 # converse relationship name is preferred if it exists.
+                 my $converse = $relationship->{converse} || $relationshipName;
+                 push @retVal, $converse;
+             }
+         }
+     }
+     # Return the result.
+     return @retVal;
+ }
+ =head3 GetDataTypes
+     my %types = ERDB::GetDataTypes();
+ Return a table of ERDB data types. The table returned is a hash of hashes.
+ The keys of the big hash are the datatypes. Each smaller hash has several
+ values used to manage the data. The most interesting is the SQL type (key
+ C<sqlType>) and the descriptive node (key C<notes>).
+ Note that changing the values in the smaller hashes will seriously break
+ things, so this data should be treated as read-only.
+ =cut
+ sub GetDataTypes {
+     return %TypeTable;
+ }
+ =head3 IsEntity
+     my $flag = $erdb->IsEntity($entityName);
+ Return TRUE if the parameter is an entity name, else FALSE.
+ =over 4
+ =item entityName
+ Object name to be tested.
+ =item RETURN
+ Returns TRUE if the specified string is an entity name, else FALSE.
+ =back
+ =cut
+ sub IsEntity {
+     # Get the parameters.
+     my ($self, $entityName) = @_;
+     # Test to see if it's an entity.
+     return exists $self->{_metaData}->{Entities}->{$entityName};
+ }
  =head3 Get
- C<< my $query = $database->Get(\@objectNames, $filterClause, $param1, $param2, ..., $paramN); >>
+     my $query = $erdb->Get(\@objectNames, $filterClause, \@params);
  This method returns a query object for entities of a specified type using a specified filter.
  The filter is a standard WHERE/ORDER BY clause with question marks as parameter markers and each
-Line 779
+Line 1654
  following call requests all B<Genome> objects for the genus specified in the variable
  $genus.
- C<< $query = $sprout->Get(['Genome'], "Genome(genus) = ?", $genus); >>
+     $query = $erdb->Get(['Genome'], "Genome(genus) = ?", [$genus]);
  The WHERE clause contains a single question mark, so there is a single additional
  parameter representing the parameter value. It would also be possible to code
- C<< $query = $sprout->Get(['Genome'], "Genome(genus) = \'$genus\'"); >>
+     $query = $erdb->Get(['Genome'], "Genome(genus) = \'$genus\'");
  however, this version of the call would generate a syntax error if there were any quote
  characters inside the variable C<$genus>.
-Line 796
+Line 1671
  It is possible to specify multiple entity and relationship names in order to retrieve more than
  one object's data at the same time, which allows highly complex joined queries. For example,
- C<< $query = $sprout->Get(['Genome', 'ComesFrom', 'Source'], "Genome(genus) = ?", $genus); >>
+     $query = $erdb->Get(['Genome', 'ComesFrom', 'Source'], "Genome(genus) = ?", [$genus]);
  If multiple names are specified, then the query processor will automatically determine a
  join path between the entities and relationships. The algorithm used is very simplistic.
- In particular, you can't specify any entity or relationship more than once, and if a
+ In particular, if a relationship is recursive, the path is determined by the order in which
- relationship is recursive, the path is determined by the order in which the entity
+ the entity and the relationship appear. For example, consider a recursive relationship
- and the relationship appear. For example, consider a recursive relationship B<IsParentOf>
+ B<IsParentOf> which relates B<People> objects to other B<People> objects. If the join path is
- which relates B<People> objects to other B<People> objects. If the join path is
  coded as C<['People', 'IsParentOf']>, then the people returned will be parents. If, however,
  the join path is C<['IsParentOf', 'People']>, then the people returned will be children.
+ If an entity or relationship is mentioned twice, the name for the second occurrence will
+ be suffixed with C<2>, the third occurrence will be suffixed with C<3>, and so forth. So,
+ for example, if we have C<['Feature', 'HasContig', 'Contig', 'HasContig']>, then the
+ B<to-link> field of the first B<HasContig> is specified as C<HasContig(to-link)>, while
+ the B<to-link> field of the second B<HasContig> is specified as C<HasContig2(to-link)>.
  =over 4
  =item objectNames
-Line 827
+Line 1707
  with an ORDER BY clause. For example, the following filter string gets all genomes for a
  particular genus and sorts them by species name.
- C<< "Genome(genus) = ? ORDER BY Genome(species)" >>
+     "Genome(genus) = ? ORDER BY Genome(species)"
+ Note that the case is important. Only an uppercase "ORDER BY" with a single space will
+ be processed. The idea is to make it less likely to find the verb by accident.
  The rules for field references in a sort order are the same as those for field references in the
  filter clause in general; however, odd things may happen if a sort field is from a secondary
  relation.
- =item param1, param2, ..., paramN
+ Finally, you can limit the number of rows returned by adding a LIMIT clause. The LIMIT must
+ be the last thing in the filter clause, and it contains only the word "LIMIT" followed by
+ a positive number. So, for example
+     "Genome(genus) = ? ORDER BY Genome(species) LIMIT 10"
+ will only return the first ten genomes for the specified genus. The ORDER BY clause is not
+ required. For example, to just get the first 10 genomes in the B<Genome> table, you could
+ use
+     "LIMIT 10"
- Parameter values to be substituted into the filter clause.
+ =item params
+ Reference to a list of parameter values to be substituted into the filter clause.
  =item RETURN
-Line 847
+Line 1742
  sub Get {
      # Get the parameters.
-     my ($self, $objectNames, $filterClause, @params) = @_;
+     my ($self, $objectNames, $filterClause, $params) = @_;
-     # Construct the SELECT statement. The general pattern is
+     # Process the SQL stuff.
-     #
+     my ($suffix, $mappedNameListRef, $mappedNameHashRef) =
-     # SELECT name1.*, name2.*, ... nameN.* FROM name1, name2, ... nameN
+         $self->_SetupSQL($objectNames, $filterClause);
-     #
+     # Create the query.
-     my $dbh = $self->{_dbh};
+     my $command = "SELECT " . join(".*, ", @{$mappedNameListRef}) .
-     my $command = "SELECT DISTINCT " . join('.*, ', @{$objectNames}) . ".* FROM " .
+         ".* $suffix";
-                 join(', ', @{$objectNames});
+     my $sth = $self->_GetStatementHandle($command, $params);
-     # Check for a filter clause.
+     # Now we create the relation map, which enables DBQuery to determine the order, name
-     if ($filterClause) {
+     # and mapped name for each object in the query.
-         # Here we have one, so we convert its field names and add it to the query. First,
+     my @relationMap = ();
-         # We create a copy of the filter string we can work with.
+     for my $mappedName (@{$mappedNameListRef}) {
-         my $filterString = $filterClause;
+         push @relationMap, [$mappedName, $mappedNameHashRef->{$mappedName}];
-         # Next, we sort the object names by length. This helps protect us from finding
-         # object names inside other object names when we're doing our search and replace.
-         my @sortedNames = sort { length($b) - length($a) } @{$objectNames};
-         # We will also keep a list of conditions to add to the WHERE clause in order to link
-         # entities and relationships as well as primary relations to secondary ones.
-         my @joinWhere = ();
-         # The final preparatory step is to create a hash table of relation names. The
-         # table begins with the relation names already in the SELECT command.
-         my %fromNames = ();
-         for my $objectName (@sortedNames) {
-             $fromNames{$objectName} = 1;
-         }
-         # We are ready to begin. We loop through the object names, replacing each
-         # object name's field references by the corresponding SQL field reference.
-         # Along the way, if we find a secondary relation, we will need to add it
-         # to the FROM clause.
-         for my $objectName (@sortedNames) {
-             # Get the length of the object name plus 2. This is the value we add to the
-             # size of the field name to determine the size of the field reference as a
-             # whole.
-             my $nameLength = 2 + length $objectName;
-             # Get the object's field list.
-             my $fieldList = $self->_GetFieldTable($objectName);
-             # Find the field references for this object.
-             while ($filterString =~ m/$objectName\(([^)]*)\)/g) {
-                 # At this point, $1 contains the field name, and the current position
-                 # is set immediately after the final parenthesis. We pull out the name of
-                 # the field and the position and length of the field reference as a whole.
-                 my $fieldName = $1;
-                 my $len = $nameLength + length $fieldName;
-                 my $pos = pos($filterString) - $len;
-                 # Insure the field exists.
-                 if (!exists $fieldList->{$fieldName}) {
-                     Confess("Field $fieldName not found for object $objectName.");
-                 } else {
-                     # Get the field's relation.
-                     my $relationName = $fieldList->{$fieldName}->{relation};
-                     # Insure the relation is in the FROM clause.
-                     if (!exists $fromNames{$relationName}) {
-                         # Add the relation to the FROM clause.
-                         $command .= ", $relationName";
-                         # Create its join sub-clause.
-                         push @joinWhere, "$objectName.id = $relationName.id";
-                         # Denote we have it available for future fields.
-                         $fromNames{$relationName} = 1;
-                     }
-                     # Form an SQL field reference from the relation name and the field name.
-                     my $sqlReference = "$relationName." . _FixName($fieldName);
-                     # Put it into the filter string in place of the old value.
-                     substr($filterString, $pos, $len) = $sqlReference;
-                     # Reposition the search.
-                     pos $filterString = $pos + length $sqlReference;
-                 }
-             }
-         }
-         # The next step is to join the objects together. We only need to do this if there
-         # is more than one object in the object list. We start with the first object and
-         # run through the objects after it. Note also that we make a safety copy of the
-         # list before running through it.
-         my @objectList = @{$objectNames};
-         my $lastObject = shift @objectList;
-         # Get the join table.
-         my $joinTable = $self->{_metaData}->{Joins};
-         # Loop through the object list.
-         for my $thisObject (@objectList) {
-             # Look for a join.
-             my $joinKey = "$lastObject/$thisObject";
-             if (!exists $joinTable->{$joinKey}) {
-                 # Here there's no join, so we throw an error.
-                 Confess("No join exists to connect from $lastObject to $thisObject.");
-             } else {
-                 # Get the join clause and add it to the WHERE list.
-                 push @joinWhere, $joinTable->{$joinKey};
-                 # Save this object as the last object for the next iteration.
-                 $lastObject = $thisObject;
-             }
-         }
-         # Now we need to handle the whole ORDER BY thing. We'll put the order by clause
-         # in the following variable.
-         my $orderClause = "";
-         # Locate the ORDER BY verb (if any).
-         if ($filterString =~ m/^(.*)ORDER BY/g) {
-             # Here we have an ORDER BY verb. Split it off of the filter string.
-             my $pos = pos $filterString;
-             $orderClause = substr($filterString, $pos);
-             $filterString = $1;
-         }
-         # Add the filter and the join clauses (if any) to the SELECT command.
-         if ($filterString) {
-             push @joinWhere, "($filterString)";
-         }
-         if (@joinWhere) {
-             $command .= " WHERE " . join(' AND ', @joinWhere);
-         }
-         # Add the sort clause (if any) to the SELECT command.
-         if ($orderClause) {
-             $command .= " ORDER BY $orderClause";
-         }
      }
-     Trace("SQL query: $command") if T(2);
-     Trace("PARMS: '" . (join "', '", @params) . "'") if (T(3) && (@params > 0));
-     my $sth = $dbh->prepare_command($command);
-     # Execute it with the parameters bound in.
-     $sth->execute(@params) || Confess("SELECT error" . $sth->errstr());
      # Return the statement object.
-     my $retVal = DBQuery::_new($self, $sth, @{$objectNames});
+     my $retVal = DBQuery::_new($self, $sth, \@relationMap);
      return $retVal;
  }
- =head3 GetList
- C<< my @dbObjects = $database->GetList(\@objectNames, $filterClause, $param1, $param2, ..., $paramN); >>
- Return a list of object descriptors for the specified objects as determined by the
+ =head3 Search
- specified filter clause.
- This method is essentially the same as L</Get> except it returns a list of objects rather
+     my $query = $erdb->Search($searchExpression, $idx, \@objectNames, $filterClause, \@params);
- than a query object that can be used to get the results one record at a time.
+ Perform a full text search with filtering. The search will be against a specified object
+ in the object name list. That object will get an extra field containing the search
+ relevance. Note that except for the search expression, the parameters of this method are
+ the same as those for L</Get> and follow the same rules.
  =over 4
+ =item searchExpression
+ Boolean search expression for the text fields of the target object. The default mode for
+ a Boolean search expression is OR, but we want the default to be AND, so we will
+ add a C<+> operator to each word with no other operator before it.
+ =item idx
+ Index in the I<$objectNames> list of the table to be searched in full-text mode.
  =item objectNames
  List containing the names of the entity and relationship objects to be retrieved.
-Line 995
+Line 1798
  or secondary entity relations; however, all of the entities and relationships involved must
  be included in the list of object names.
- The filter clause can also specify a sort order. To do this, simply follow the filter string
+ =item params
- with an ORDER BY clause. For example, the following filter string gets all genomes for a
- particular genus and sorts them by species name.
- C<< "Genome(genus) = ? ORDER BY Genome(species)" >>
- The rules for field references in a sort order are the same as those for field references in the
- filter clause in general; however, odd things may happen if a sort field is from a secondary
- relation.
- =item param1, param2, ..., paramN
- Parameter values to be substituted into the filter clause.
+ Reference to a list of parameter values to be substituted into the filter clause.
  =item RETURN
- Returns a list of B<DBObject>s that satisfy the query conditions.
+ Returns a query object for the specified search.
  =back
  =cut
- #: Return Type @%
- sub GetList {
+ sub Search {
      # Get the parameters.
-     my ($self, $objectNames, $filterClause, @params) = @_;
+     my ($self, $searchExpression, $idx, $objectNames, $filterClause, $params) = @_;
      # Declare the return variable.
-     my @retVal = ();
+     my $retVal;
-     # Perform the query.
+     # Create a safety copy of the parameter list. Note we have to be careful to insure
-     my $query = $self->Get($objectNames, $filterClause, @params);
+     # a parameter list exists before we copy it.
-     # Loop through the results.
+     my @myParams = ();
-     while (my $object = $query->Fetch) {
+     if (defined $params) {
-         push @retVal, $object;
+         @myParams = @{$params};
+     }
+     # Get the first object's structure so we have access to the searchable fields.
+     my $object1Name = $objectNames->[$idx];
+     my $object1Structure = $self->_GetStructure($object1Name);
+     # Get the field list.
+     if (! exists $object1Structure->{searchFields}) {
+         Confess("No searchable index for $object1Name.");
+     } else {
+         # Get the field list.
+         my @fields = @{$object1Structure->{searchFields}};
+         # Clean the search expression.
+         my $actualKeywords = $self->CleanKeywords($searchExpression);
+         # Prefix a "+" to each uncontrolled word. This converts the default
+         # search mode from OR to AND.
+         $actualKeywords =~ s/(^|\s)(\w|")/$1\+$2/g;
+         Trace("Actual keywords for search are\n$actualKeywords") if T(3);
+         # We need two match expressions, one for the filter clause and one in the
+         # query itself. Both will use a parameter mark, so we need to push the
+         # search expression onto the front of the parameter list twice.
+         unshift @myParams, $actualKeywords, $actualKeywords;
+         # Build the match expression.
+         my @matchFilterFields = map { "$object1Name." . _FixName($_) } @fields;
+         my $matchClause = "MATCH (" . join(", ", @matchFilterFields) . ") AGAINST (? IN BOOLEAN MODE)";
+         # Process the SQL stuff.
+         my ($suffix, $mappedNameListRef, $mappedNameHashRef) =
+             $self->_SetupSQL($objectNames, $filterClause, $matchClause);
+         # Create the query. Note that the match clause is inserted at the front of
+         # the select fields.
+         my $command = "SELECT $matchClause, " . join(".*, ", @{$mappedNameListRef}) .
+             ".* $suffix";
+         my $sth = $self->_GetStatementHandle($command, \@myParams);
+         # Now we create the relation map, which enables DBQuery to determine the order, name
+         # and mapped name for each object in the query.
+         my @relationMap = _RelationMap($mappedNameHashRef, $mappedNameListRef);
+         # Return the statement object.
+         $retVal = DBQuery::_new($self, $sth, \@relationMap, $object1Name);
      }
-     # Return the result.
+     return $retVal;
-     return @retVal;
  }
- =head3 ComputeObjectSentence
+ =head3 GetFlat
- C<< my $sentence = $database->ComputeObjectSentence($objectName); >>
+     my @list = $erdb->GetFlat(\@objectNames, $filterClause, \@parameterList, $field);
- Check an object name, and if it is a relationship convert it to a relationship sentence.
+ This is a variation of L</GetAll> that asks for only a single field per record and
+ returns a single flattened list.
  =over 4
- =item objectName
+ =item objectNames
- Name of the entity or relationship.
+ List containing the names of the entity and relationship objects to be retrieved.
+ =item filterClause
+ WHERE/ORDER BY clause (without the WHERE) to be used to filter and sort the query. The WHERE clause can
+ be parameterized with parameter markers (C<?>). Each field used must be specified in the standard form
+ B<I<objectName>(I<fieldName>)>. Any parameters specified in the filter clause should be added to the
+ parameter list as additional parameters. The fields in a filter clause can come from primary
+ entity relations, relationship relations, or secondary entity relations; however, all of the
+ entities and relationships involved must be included in the list of object names.
+ =item parameterList
+ List of the parameters to be substituted in for the parameters marks in the filter clause.
+ =item field
+ Name of the field to be used to get the elements of the list returned.
  =item RETURN
- Returns a string containing the entity name or a relationship sentence.
+ Returns a list of values.
  =back
  =cut
+ #: Return Type @;
- sub ComputeObjectSentence {
+ sub GetFlat {
      # Get the parameters.
-     my ($self, $objectName) = @_;
+     my ($self, $objectNames, $filterClause, $parameterList, $field) = @_;
-     # Set the default return value.
+     # Construct the query.
-     my $retVal = $objectName;
+     my $query = $self->Get($objectNames, $filterClause, $parameterList);
-     # Look for the object as a relationship.
+     # Create the result list.
-     my $relTable = $self->{_metaData}->{Relationships};
+     my @retVal = ();
-     if (exists $relTable->{$objectName}) {
+     # Loop through the records, adding the field values found to the result list.
-         # Get the relationship sentence.
+     while (my $row = $query->Fetch()) {
-         $retVal = _ComputeRelationshipSentence($objectName, $relTable->{$objectName});
+         push @retVal, $row->Value($field);
      }
-     # Return the result.
+     # Return the list created.
-     return $retVal;
+     return @retVal;
  }
- =head3 DumpRelations
+ =head3 SpecialFields
- C<< $database->DumpRelations($outputDirectory); >>
+     my %specials = $erdb->SpecialFields($entityName);
- Write the contents of all the relations to tab-delimited files in the specified directory.
+ Return a hash mapping special fields in the specified entity to the value of their
- Each file will have the same name as the relation dumped, with an extension of DTX.
+ C<special> attribute. This enables the subclass to get access to the special field
+ attributes without needed to plumb the internal ERDB data structures.
  =over 4
- =item outputDirectory
+ =item entityName
- Name of the directory into which the relation files should be dumped.
+ Name of the entity whose special fields are desired.
+ =item RETURN
+ Returns a hash. The keys of the hash are the special field names, and the values
+ are the values from each special field's C<special> attribute.
  =back
  =cut
- sub DumpRelations {
+ sub SpecialFields {
      # Get the parameters.
-     my ($self, $outputDirectory) = @_;
+     my ($self, $entityName) = @_;
-     # Now we need to run through all the relations. First, we loop through the entities.
+     # Declare the return variable.
-     my $metaData = $self->{_metaData};
+     my %retVal = ();
-     my $entities = $metaData->{Entities};
+     # Find the entity's data structure.
-     for my $entityName (keys %{$entities}) {
+     my $entityData = $self->{_metaData}->{Entities}->{$entityName};
-         my $entityStructure = $entities->{$entityName};
+     # Loop through its fields, adding each special field to the return hash.
-         # Get the entity's relations.
+     my $fieldHash = $entityData->{Fields};
-         my $relationList = $entityStructure->{Relations};
+     for my $fieldName (keys %{$fieldHash}) {
-         # Loop through the relations, dumping them.
+         my $fieldData = $fieldHash->{$fieldName};
-         for my $relationName (keys %{$relationList}) {
+         if (exists $fieldData->{special}) {
-             my $relation = $relationList->{$relationName};
+             $retVal{$fieldName} = $fieldData->{special};
-             $self->_DumpRelation($outputDirectory, $relationName, $relation);
-         }
      }
-     # Next, we loop through the relationships.
-     my $relationships = $metaData->{Relationships};
-     for my $relationshipName (keys %{$relationships}) {
-         my $relationshipStructure = $relationships->{$relationshipName};
-         # Dump this relationship's relation.
-         $self->_DumpRelation($outputDirectory, $relationshipName, $relationshipStructure->{Relations}->{$relationshipName});
      }
+     # Return the result.
+     return %retVal;
  }
- =head3 InsertObject
+ =head3 Delete
- C<< my $ok = $database->InsertObject($objectType, \%fieldHash); >>
+     my $stats = $erdb->Delete($entityName, $objectID, %options);
- Insert an object into the database. The object is defined by a type name and then a hash
+ Delete an entity instance from the database. The instance is deleted along with all entity and
- of field names to values. Field values in the primary relation are represented by scalars.
+ relationship instances dependent on it. The definition of I<dependence> is recursive.
- (Note that for relationships, the primary relation is the B<only> relation.)
- Field values for the other relations comprising the entity are always list references. For
- example, the following line inserts an inactive PEG feature named C<fig|188.1.peg.1> with aliases
- C<ZP_00210270.1> and C<gi|46206278>.
- C<< $database->InsertObject('Feature', { id => 'fig|188.1.peg.1', active => 0, feature-type => 'peg', alias => ['ZP_00210270.1', 'gi|46206278']}); >>
+ An object is always dependent on itself. An object is dependent if it is a 1-to-many or many-to-many
+ relationship connected to a dependent entity or if it is the "to" entity connected to a 1-to-many
+ dependent relationship.
- The next statement inserts a C<HasProperty> relationship between feature C<fig|158879.1.peg.1> and
+ =over 4
- property C<4> with an evidence URL of C<http://seedu.uchicago.edu/query.cgi?article_id=142>.
- C<< $database->InsertObject('HasProperty', { 'from-link' => 'fig|158879.1.peg.1', 'to-link' => 4, evidence = 'http://seedu.uchicago.edu/query.cgi?article_id=142'}); >>
+ =item entityName
- =over 4
+ Name of the entity type for the instance being deleted.
- =item newObjectType
+ =item objectID
- Type name of the object to insert.
+ ID of the entity instance to be deleted. If the ID contains a wild card character (C<%>),
+ then it is presumed to by a LIKE pattern.
- =item fieldHash
+ =item options
- Hash of field names to values.
+ A hash detailing the options for this delete operation.
  =item RETURN
- Returns 1 if successful, 0 if an error occurred.
+ Returns a statistics object indicating how many records of each particular table were
+ deleted.
  =back
- =cut
+ The permissible options for this method are as follows.
- sub InsertObject {
+ =over 4
+ =item testMode
+ If TRUE, then the delete statements will be traced, but no changes will be made to the database.
+ =item keepRoot
+ If TRUE, then the entity instances will not be deleted, only the dependent records.
+ =back
+ =cut
+ #: Return Type $%;
+ sub Delete {
      # Get the parameters.
-     my ($self, $newObjectType, $fieldHash) = @_;
+     my ($self, $entityName, $objectID, %options) = @_;
-     # Denote that so far we appear successful.
+     # Declare the return variable.
-     my $retVal = 1;
+     my $retVal = Stats->new();
-     # Get the database handle.
+     # Get the DBKernel object.
-     my $dbh = $self->{_dbh};
+     my $db = $self->{_dbh};
-     # Get the relation list.
+     # We're going to generate all the paths branching out from the starting entity. One of
-     my $relationTable = $self->_GetRelationTable($newObjectType);
+     # the things we have to be careful about is preventing loops. We'll use a hash to
-     # Loop through the relations. We'll build insert statements for each one. If a relation is
+     # determine if we've hit a loop.
-     # secondary, we may end up generating multiple insert statements. If an error occurs, we
+     my %alreadyFound = ();
-     # stop the loop.
+     # These next lists will serve as our result stack. We start by pushing object lists onto
-     my @relationList = keys %{$relationTable};
+     # the stack, and then popping them off to do the deletes. This means the deletes will
-     for (my $i = 0; $retVal && $i <= $#relationList; $i++) {
+     # start with the longer paths before getting to the shorter ones. That, in turn, makes
-         my $relationName = $relationList[$i];
+     # sure we don't delete records that might be needed to forge relationships back to the
-         my $relationDefinition = $relationTable->{$relationName};
+     # original item. We have two lists-- one for TO-relationships, and one for
-         # Get the relation's fields. For each field we will collect a value in the corresponding
+     # FROM-relationships and entities.
-         # position of the @valueList array. If one of the fields is missing, we will add it to the
+     my @fromPathList = ();
-         # @missing list.
+     my @toPathList = ();
-         my @fieldList = @{$relationDefinition->{Fields}};
+     # This final list is used to remember what work still needs to be done. We push paths
-         my @fieldNameList = ();
+     # onto the list, then pop them off to extend the paths. We prime it with the starting
-         my @valueList = ();
+     # point. Note that we will work hard to insure that the last item on a path in the
-         my @missing = ();
+     # to-do list is always an entity.
-         my $recordCount = 1;
+     my @todoList = ([$entityName]);
-         for my $fieldDescriptor (@fieldList) {
+     while (@todoList) {
-             # Get the field name and save it. Note we need to fix it up so the hyphens
+         # Get the current path.
-             # are converted to underscores.
+         my $current = pop @todoList;
-             my $fieldName = $fieldDescriptor->{name};
+         # Copy it into a list.
-             push @fieldNameList, _FixName($fieldName);
+         my @stackedPath = @{$current};
-             # Look for the named field in the incoming structure. Note that we are looking
+         # Pull off the last item on the path. It will always be an entity.
-             # for the real field name, not the fixed-up one!
+         my $myEntityName = pop @stackedPath;
-             if (exists $fieldHash->{$fieldName}) {
+         # Add it to the alreadyFound list.
-                 # Here we found the field. Stash it in the value list.
+         $alreadyFound{$myEntityName} = 1;
-                 my $value = $fieldHash->{$fieldName};
+         # Figure out if we need to delete this entity.
-                 push @valueList, $value;
+         if ($myEntityName ne $entityName || ! $options{keepRoot}) {
-                 # If the value is a list, we may need to increment the record count.
+             # Get the entity data.
-                 if (ref $value eq "ARRAY") {
+             my $entityData = $self->_GetStructure($myEntityName);
-                     my $thisCount = @{$value};
+             # Loop through the entity's relations. A DELETE command will be needed for each of them.
-                     if ($recordCount == 1) {
+             my $relations = $entityData->{Relations};
-                         # Here we have our first list, so we save its count.
+             for my $relation (keys %{$relations}) {
-                         $recordCount = $thisCount;
+                 my @augmentedList = (@stackedPath, $relation);
-                     } elsif ($recordCount != $thisCount) {
+                 push @fromPathList, \@augmentedList;
-                         # Here we have a second list, so its length has to match the
-                         # previous lists.
-                         Trace("Field $value in new $newObjectType object has an invalid list length $thisCount. Expected $recordCount.") if T(0);
-                         $retVal = 0;
                      }
                  }
+         # Now we need to look for relationships connected to this entity.
+         my $relationshipList = $self->{_metaData}->{Relationships};
+         for my $relationshipName (keys %{$relationshipList}) {
+             my $relationship = $relationshipList->{$relationshipName};
+             # Check the FROM field. We're only interested if it's us.
+             if ($relationship->{from} eq $myEntityName) {
+                 # Add the path to this relationship.
+                 my @augmentedList = (@stackedPath, $myEntityName, $relationshipName);
+                 push @fromPathList, \@augmentedList;
+                 # Check the arity. If it's MM we're done. If it's 1M
+                 # and the target hasn't been seen yet, we want to
+                 # stack the entity for future processing.
+                 if ($relationship->{arity} eq '1M') {
+                     my $toEntity = $relationship->{to};
+                     if (! exists $alreadyFound{$toEntity}) {
+                         # Here we have a new entity that's dependent on
+                         # the current entity, so we need to stack it.
+                         my @stackList = (@augmentedList, $toEntity);
+                         push @fromPathList, \@stackList;
              } else {
-                 # Here the field is not present. Flag it as missing.
+                         Trace("$toEntity ignored because it occurred previously.") if T(4);
-                 push @missing, $fieldName;
              }
          }
-         # If we are the primary relation, add the new-record flag.
-         if ($relationName eq $newObjectType) {
-             push @valueList, 1;
-             push @fieldNameList, "new_record";
          }
-         # Only proceed if there are no missing fields.
+             # Now check the TO field. In this case only the relationship needs
-         if (@missing > 0) {
+             # deletion.
-             Trace("Relation $relationName for $newObjectType skipped due to missing fields: " .
+             if ($relationship->{to} eq $myEntityName) {
-                 join(' ', @missing)) if T(1);
+                 my @augmentedList = (@stackedPath, $myEntityName, $relationshipName);
-         } else {
+                 push @toPathList, \@augmentedList;
-             # Build the INSERT statement.
-             my $statement = "INSERT INTO $relationName (" . join (', ', @fieldNameList) .
-                 ") VALUES (";
-             # Create a marker list of the proper size and put it in the statement.
-             my @markers = ();
-             while (@markers < @fieldNameList) { push @markers, '?'; }
-             $statement .= join(', ', @markers) . ")";
-             # We have the insert statement, so prepare it.
-             my $sth = $dbh->prepare_command($statement);
-             Trace("Insert statement prepared: $statement") if T(3);
-             # Now we loop through the values. If a value is scalar, we use it unmodified. If it's
-             # a list, we use the current element. The values are stored in the @parameterList array.
-             my $done = 0;
-             for (my $i = 0; $i < $recordCount; $i++) {
-                 # Clear the parameter list array.
-                 my @parameterList = ();
-                 # Loop through the values.
-                 for my $value (@valueList) {
-                     # Check to see if this is a scalar value.
-                     if (ref $value eq "ARRAY") {
-                         # Here we have a list value. Pull the current entry.
-                         push @parameterList, $value->[$i];
-                     } else {
-                         # Here we have a scalar value. Use it unmodified.
-                         push @parameterList, $value;
                      }
                  }
-                 # Execute the INSERT statement with the specified parameter list.
-                 $retVal = $sth->execute(@parameterList);
-                 if (!$retVal) {
-                     my $errorString = $sth->errstr();
-                     Trace("Insert error: $errorString.") if T(0);
                  }
+     # Create the first qualifier for the WHERE clause. This selects the
+     # keys of the primary entity records to be deleted. When we're deleting
+     # from a dependent table, we construct a join path from the first qualifier
+     # to the table containing the dependent records to delete.
+     my $qualifier = ($objectID =~ /%/ ? "LIKE ?" : "= ?");
+     # We need to make two passes. The first is through the to-list, and
+     # the second through the from-list. The from-list is second because
+     # the to-list may need to pass through some of the entities the
+     # from-list would delete.
+     my %stackList = ( from_link => \@fromPathList, to_link => \@toPathList );
+     # Now it's time to do the deletes. We do it in two passes.
+     for my $keyName ('to_link', 'from_link') {
+         # Get the list for this key.
+         my @pathList = @{$stackList{$keyName}};
+         Trace(scalar(@pathList) . " entries in path list for $keyName.") if T(3);
+         # Loop through this list.
+         while (my $path = pop @pathList) {
+             # Get the table whose rows are to be deleted.
+             my @pathTables = @{$path};
+             # Start the DELETE statement. We need to call DBKernel because the
+             # syntax of a DELETE-USING varies among DBMSs.
+             my $target = $pathTables[$#pathTables];
+             my $stmt = $db->SetUsing(@pathTables);
+             # Now start the WHERE. The first thing is the ID field from the starting table. That
+             # starting table will either be the entity relation or one of the entity's
+             # sub-relations.
+             $stmt .= " WHERE $pathTables[0].id $qualifier";
+             # Now we run through the remaining entities in the path, connecting them up.
+             for (my $i = 1; $i <= $#pathTables; $i += 2) {
+                 # Connect the current relationship to the preceding entity.
+                 my ($entity, $rel) = @pathTables[$i-1,$i];
+                 # The style of connection depends on the direction of the relationship.
+                 $stmt .= " AND $entity.id = $rel.$keyName";
+                 if ($i + 1 <= $#pathTables) {
+                     # Here there's a next entity, so connect that to the relationship's
+                     # to-link.
+                     my $entity2 = $pathTables[$i+1];
+                     $stmt .= " AND $rel.to_link = $entity2.id";
+                 }
+             }
+             # Now we have our desired DELETE statement.
+             if ($options{testMode}) {
+                 # Here the user wants to trace without executing.
+                 Trace($stmt) if T(0);
+             } else {
+                 # Here we can delete. Note that the SQL method dies with a confession
+                 # if an error occurs, so we just go ahead and do it.
+                 Trace("Executing delete from $target using '$objectID'.") if T(3);
+                 my $rv = $db->SQL($stmt, 0, $objectID);
+                 # Accumulate the statistics for this delete. The only rows deleted
+                 # are from the target table, so we use its name to record the
+                 # statistic.
+                 $retVal->Add($target, $rv);
              }
          }
      }
-     # Return the success indicator.
+     # Return the result.
      return $retVal;
  }
- =head3 LoadTable
+ =head3 Disconnect
- C<< my %results = $database->LoadTable($fileName, $relationName, $truncateFlag); >>
+     $erdb->Disconnect($relationshipName, $originEntityName, $originEntityID);
- Load data from a tab-delimited file into a specified table, optionally re-creating the table
+ Disconnect an entity instance from all the objects to which it is related. This
- first.
+ will delete each relationship instance that connects to the specified entity.
  =over 4
- =item fileName
+ =item relationshipName
- Name of the file from which the table data should be loaded.
- =item relationName
- Name of the relation to be loaded. This is the same as the table name.
+ Name of the relationship whose instances are to be deleted.
- =item truncateFlag
+ =item originEntityName
- TRUE if the table should be dropped and re-created, else FALSE
+ Name of the entity that is to be disconnected.
- =item RETURN
+ =item originEntityID
- Returns a statistical object containing the number of records read and a list of
+ ID of the entity that is to be disconnected.
- the error messages.
  =back
  =cut
- sub LoadTable {
+ sub Disconnect {
      # Get the parameters.
-     my ($self, $fileName, $relationName, $truncateFlag) = @_;
+     my ($self, $relationshipName, $originEntityName, $originEntityID) = @_;
-     # Create the statistical return object.
+     # Get the relationship descriptor.
-     my $retVal = _GetLoadStats();
+     my $structure = $self->_GetStructure($relationshipName);
-     # Trace the fact of the load.
+     # Insure we have a relationship.
-     Trace("Loading table $relationName from $fileName") if T(2);
+     if (! exists $structure->{from}) {
+         Confess("$relationshipName is not a relationship in the database.");
+     } else {
      # Get the database handle.
      my $dbh = $self->{_dbh};
-     # Get the relation data.
+         # We'll set this value to 1 if we find our entity.
-     my $relation = $self->_FindRelation($relationName);
+         my $found = 0;
-     # Check the truncation flag.
+         # Loop through the ends of the relationship.
-     if ($truncateFlag) {
+         for my $dir ('from', 'to') {
-         Trace("Creating table $relationName") if T(2);
+             if ($structure->{$dir} eq $originEntityName) {
-         # Re-create the table without its index.
+                 $found = 1;
-         $self->CreateTable($relationName, 0);
+                 # Here we want to delete all relationship instances on this side of the
-         # If this is a pre-index DBMS, create the index here.
+                 # entity instance.
-         if ($dbh->{_preIndex}) {
+                 Trace("Disconnecting in $dir direction with ID \"$originEntityID\".");
-             eval {
+                 # We do this delete in batches to keep it from dragging down the
-                 $self->CreateIndex($relationName);
+                 # server.
-             };
+                 my $limitClause = ($FIG_Config::delete_limit ? "LIMIT $FIG_Config::delete_limit" : "");
-             if ($@) {
+                 my $done = 0;
-                 $retVal->AddMessage($@);
+                 while (! $done) {
-             }
+                     # Do the delete.
+                     my $rows = $dbh->SQL("DELETE FROM $relationshipName WHERE ${dir}_link = ? $limitClause", 0, $originEntityID);
+                     # See if we're done. We're done if no rows were found or the delete is unlimited.
+                     $done = ($rows == 0 || ! $limitClause);
          }
      }
-     # Determine whether or not this is a primary relation. Primary relations have an extra
-     # field indicating whether or not a given object is new or was loaded from the flat files.
-     my $primary = $self->_IsPrimary($relationName);
-     # Get the number of fields in this relation.
-     my @fieldList = @{$relation->{Fields}};
-     my $fieldCount = @fieldList;
-     # Start a database transaction.
-     $dbh->begin_tran;
-     # Open the relation file. We need to create a cleaned-up copy before loading.
-     open TABLEIN, '<', $fileName;
-     my $tempName = "$fileName.tbl";
-     open TABLEOUT, '>', $tempName;
-     my $inputCount = 0;
-     # Loop through the file.
-     while (<TABLEIN>) {
-         $inputCount++;
-         # Chop off the new-line character.
-         my $record = Tracer::Strip($_);
-         # Only proceed if the record is non-blank.
-         if ($record) {
-             # Escape all the backslashes found in the line.
-             $record =~ s/\\/\\\\/g;
-             # Insure the number of fields is correct.
-             my @fields = split /\t/, $record;
-             while (@fields > $fieldCount) {
-                 my $extraField = $fields[$#fields];
-                 delete $fields[$#fields];
-                 if ($extraField) {
-                     Trace("Nonblank extra field value \"$extraField\" deleted from record $inputCount of $fileName.") if T(1);
-                 }
-             }
-             while (@fields < $fieldCount) {
-                 push @fields, "";
-             }
-             # If this is a primary relation, add a 0 for the new-record flag (indicating that
-             # this record is not new, but part of the original load).
-             if ($primary) {
-                 push @fields, "0";
-             }
-             # Write the record.
-             $record = join "\t", @fields;
-             print TABLEOUT "$record\n";
-             # Count the record written.
-             my $count = $retVal->Add('records');
-             my $len = length $record;
-             Trace("Record $count written with $len characters.") if T(4);
-         } else {
-             # Here we have a blank record.
-             $retVal->Add('skipped');
-         }
-     }
-     # Close the files.
-     close TABLEIN;
-     close TABLEOUT;
-     Trace("Temporary file $tempName created.") if T(2);
-     # Load the table.
-     my $rv;
-     eval {
-         $rv = $dbh->load_table(file => $tempName, tbl => $relationName);
-     };
-     if (!defined $rv) {
-         $retVal->AddMessage($@) if ($@);
-         $retVal->AddMessage("Table load failed for $relationName using $tempName.");
-         Trace("Table load failed for $relationName.") if T(1);
-     } else {
-         # Here we successfully loaded the table. Trace the number of records loaded.
-         Trace("$retVal->{records} records read for $relationName.") if T(2);
-         # If we're rebuilding, we need to create the table indexes.
-         if ($truncateFlag && ! $dbh->{_preIndex}) {
-             eval {
-                 $self->CreateIndex($relationName);
-             };
-             if ($@) {
-                 $retVal->AddMessage($@);
              }
+         # Insure we found the entity on at least one end.
+         if (! $found) {
+             Confess("Entity \"$originEntityName\" does not use $relationshipName.");
          }
-         # Analyze the table to help optimize tables.
      }
-     # Commit the database changes.
-     $dbh->commit_tran;
-     $dbh->vacuum_it($relationName);
-     # Delete the temporary file.
-     unlink $tempName;
-     # Return the statistics.
-     return $retVal;
  }
- =head3 GenerateEntity
+ =head3 DeleteRow
- C<< my $fieldHash = $database->GenerateEntity($id, $type, \%values); >>
+     $erdb->DeleteRow($relationshipName, $fromLink, $toLink, \%values);
- Generate the data for a new entity instance. This method creates a field hash suitable for
+ Delete a row from a relationship. In most cases, only the from-link and to-link are
- passing as a parameter to L</InsertObject>. The ID is specified by the callr, but the rest
+ needed; however, for relationships with intersection data values can be specified
- of the fields are generated using information in the database schema.
+ for the other fields using a hash.
- Each data type has a default algorithm for generating random test data. This can be overridden
- by including a B<DataGen> element in the field. If this happens, the content of the element is
- executed as a PERL program in the context of this module. The element may make use of a C<$this>
- variable which contains the field hash as it has been built up to the current point. If any
- fields are dependent on other fields, the C<pass> attribute can be used to control the order
- in which the fields are generated. A field with a high data pass number will be generated after
- a field with a lower one. If any external values are needed, they should be passed in via the
- optional third parameter, which will be available to the data generation script under the name
- C<$value>. Several useful utility methods are provided for generating random values, including
- L</IntGen>, L</StringGen>, L</FloatGen>, and L</DateGen>. Note that dates are stored and generated
- in the form of a timestamp number rather than a string.
  =over 4
- =item id
+ =item relationshipName
- ID to assign to the new entity.
+ Name of the relationship from which the row is to be deleted.
- =item type
+ =item fromLink
+ ID of the entity instance in the From direction.
+ =item toLink
- Type name for the new entity.
+ ID of the entity instance in the To direction.
  =item values
- Hash containing additional values that might be needed by the data generation methods (optional).
+ Reference to a hash of other values to be used for filtering the delete.
  =back
  =cut
- sub GenerateEntity {
+ sub DeleteRow {
      # Get the parameters.
-     my ($self, $id, $type, $values) = @_;
+     my ($self, $relationshipName, $fromLink, $toLink, $values) = @_;
-     # Create the return hash.
+     # Create a hash of all the filter information.
-     my $this = { id => $id };
+     my %filter = ('from-link' => $fromLink, 'to-link' => $toLink);
-     # Get the metadata structure.
+     if (defined $values) {
-     my $metadata = $self->{_metaData};
+         for my $key (keys %{$values}) {
-     # Get this entity's list of fields.
+             $filter{$key} = $values->{$key};
-     if (!exists $metadata->{Entities}->{$type}) {
+         }
-         Confess("Unrecognized entity type $type in GenerateEntity.");
+     }
-     } else {
+     # Build an SQL statement out of the hash.
-         my $entity = $metadata->{Entities}->{$type};
+     my @filters = ();
-         my $fields = $entity->{Fields};
+     my @parms = ();
-         # Generate data from the fields.
+     for my $key (keys %filter) {
-         _GenerateFields($this, $fields, $type, $values);
+         push @filters, _FixName($key) . " = ?";
+         push @parms, $filter{$key};
      }
-     # Return the hash created.
+     Trace("Parms for delete row are " . join(", ", map { "\"$_\"" } @parms) . ".") if T(SQL => 4);
-     return $this;
+     my $command = "DELETE FROM $relationshipName WHERE " .
+                   join(" AND ", @filters);
+     # Execute it.
+     my $dbh = $self->{_dbh};
+     $dbh->SQL($command, undef, @parms);
  }
- =head3 GetEntity
+ =head3 DeleteLike
- C<< my $entityObject = $sprout->GetEntity($entityType, $ID); >>
+     my $deleteCount = $erdb->DeleteLike($relName, $filter, \@parms);
- Return an object describing the entity instance with a specified ID.
+ Delete all the relationship rows that satisfy a particular filter condition. Unlike a normal
+ filter, only fields from the relationship itself can be used.
  =over 4
- =item entityType
+ =item relName
+ Name of the relationship whose records are to be deleted.
+ =item filter
+ A filter clause (L</Get>-style) for the delete query.
+ =item parms
+ Reference to a list of parameters for the filter clause.
+ =item RETURN
+ Returns a count of the number of rows deleted.
+ =back
+ =cut
+ sub DeleteLike {
+     # Get the parameters.
+     my ($self, $objectName, $filter, $parms) = @_;
+     # Declare the return variable.
+     my $retVal;
+     # Insure the parms argument is an array reference if the caller left it off.
+     if (! defined($parms)) {
+         $parms = [];
+     }
+     # Insure we have a relationship. The main reason for this is if we delete an entity
+     # instance we have to yank out a bunch of other stuff with it.
+     if ($self->IsEntity($objectName)) {
+         Confess("Cannot use DeleteLike on $objectName, because it is not a relationship.");
+     } else {
+         # Create the SQL command suffix to get the desierd records.
+         my ($suffix) = $self->_SetupSQL([$objectName], $filter);
+         # Convert it to a DELETE command.
+         my $command = "DELETE $suffix";
+         # Execute the command.
+         my $dbh = $self->{_dbh};
+         my $result = $dbh->SQL($command, 0, @{$parms});
+         # Check the results. Note we convert the "0D0" result to a real zero.
+         # A failure causes an abnormal termination, so the caller isn't going to
+         # worry about it.
+         if (! defined $result) {
+             Confess("Error deleting from $objectName: " . $dbh->errstr());
+         } elsif ($result == 0) {
+             $retVal = 0;
+         } else {
+             $retVal = $result;
+         }
+     }
+     # Return the result count.
+     return $retVal;
+ }
+ =head3 SortNeeded
+     my $parms = $erdb->SortNeeded($relationName);
+ Return the pipe command for the sort that should be applied to the specified
+ relation when creating the load file.
+ For example, if the load file should be sorted ascending by the first
+ field, this method would return
+     sort -k1 -t"\t"
+ If the first field is numeric, the method would return
+     sort -k1n -t"\t"
+ Unfortunately, due to a bug in the C<sort> command, we cannot eliminate duplicate
+ keys using a sort.
+ =over 4
+ =item relationName
+ Name of the relation to be examined.
+ =item
+ Returns the sort command to use for sorting the relation, suitable for piping.
+ =back
+ =cut
+ #: Return Type $;
+ sub SortNeeded {
+     # Get the parameters.
+     my ($self, $relationName) = @_;
+     # Declare a descriptor to hold the names of the key fields.
+     my @keyNames = ();
+     # Get the relation structure.
+     my $relationData = $self->_FindRelation($relationName);
+     # Find out if the relation is a primary entity relation,
+     # a relationship relation, or a secondary entity relation.
+     my $entityTable = $self->{_metaData}->{Entities};
+     my $relationshipTable = $self->{_metaData}->{Relationships};
+     if (exists $entityTable->{$relationName}) {
+         # Here we have a primary entity relation.
+         push @keyNames, "id";
+     } elsif (exists $relationshipTable->{$relationName}) {
+         # Here we have a relationship. We sort using the FROM index.
+         my $relationshipData = $relationshipTable->{$relationName};
+         my $index = $relationData->{Indexes}->{idxFrom};
+         push @keyNames, @{$index->{IndexFields}};
+     } else {
+         # Here we have a secondary entity relation, so we have a sort on the ID field.
+         push @keyNames, "id";
+     }
+     # Now we parse the key names into sort parameters. First, we prime the return
+     # string.
+     my $retVal = "sort -S 1G -T\"$FIG_Config::temp\" -t\"\t\" ";
+     # Get the relation's field list.
+     my @fields = @{$relationData->{Fields}};
+     # Loop through the keys.
+     for my $keyData (@keyNames) {
+         # Get the key and the ordering.
+         my ($keyName, $ordering);
+         if ($keyData =~ /^([^ ]+) DESC/) {
+             ($keyName, $ordering) = ($1, "descending");
+         } else {
+             ($keyName, $ordering) = ($keyData, "ascending");
+         }
+         # Find the key's position and type.
+         my $fieldSpec;
+         for (my $i = 0; $i <= $#fields && ! $fieldSpec; $i++) {
+             my $thisField = $fields[$i];
+             if ($thisField->{name} eq $keyName) {
+                 # Get the sort modifier for this field type. The modifier
+                 # decides whether we're using a character, numeric, or
+                 # floating-point sort.
+                 my $modifier = $TypeTable{$thisField->{type}}->{sort};
+                 # If the index is descending for this field, denote we want
+                 # to reverse the sort order on this field.
+                 if ($ordering eq 'descending') {
+                     $modifier .= "r";
+                 }
+                 # Store the position and modifier into the field spec, which
+                 # will stop the inner loop. Note that the field number is
+                 # 1-based in the sort command, so we have to increment the
+                 # index.
+                 my $realI = $i + 1;
+                 $fieldSpec = "$realI,$realI$modifier";
+             }
+         }
+         # Add this field to the sort command.
+         $retVal .= " -k$fieldSpec";
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 GetList
+     my @dbObjects = $erdb->GetList(\@objectNames, $filterClause, \@params);
+ Return a list of object descriptors for the specified objects as determined by the
+ specified filter clause.
+ This method is essentially the same as L</Get> except it returns a list of objects rather
+ than a query object that can be used to get the results one record at a time.
+ =over 4
+ =item objectNames
+ List containing the names of the entity and relationship objects to be retrieved.
+ =item filterClause
+ WHERE clause (without the WHERE) to be used to filter and sort the query. The WHERE clause can
+ be parameterized with parameter markers (C<?>). Each field used in the WHERE clause must be
+ specified in the standard form B<I<objectName>(I<fieldName>)>. Any parameters specified
+ in the filter clause should be added to the parameter list as additional parameters. The
+ fields in a filter clause can come from primary entity relations, relationship relations,
+ or secondary entity relations; however, all of the entities and relationships involved must
+ be included in the list of object names.
+ The filter clause can also specify a sort order. To do this, simply follow the filter string
+ with an ORDER BY clause. For example, the following filter string gets all genomes for a
+ particular genus and sorts them by species name.
+     "Genome(genus) = ? ORDER BY Genome(species)"
+ The rules for field references in a sort order are the same as those for field references in the
+ filter clause in general; however, odd things may happen if a sort field is from a secondary
+ relation.
+ =item params
+ Reference to a list of parameter values to be substituted into the filter clause.
+ =item RETURN
+ Returns a list of B<ERDBObject>s that satisfy the query conditions.
+ =back
+ =cut
+ #: Return Type @%
+ sub GetList {
+     # Get the parameters.
+     my ($self, $objectNames, $filterClause, $params) = @_;
+     # Declare the return variable.
+     my @retVal = ();
+     # Perform the query.
+     my $query = $self->Get($objectNames, $filterClause, $params);
+     # Loop through the results.
+     while (my $object = $query->Fetch) {
+         push @retVal, $object;
+     }
+     # Return the result.
+     return @retVal;
+ }
+ =head3 GetCount
+     my $count = $erdb->GetCount(\@objectNames, $filter, \@params);
+ Return the number of rows found by a specified query. This method would
+ normally be used to count the records in a single table. For example, in a
+ genetics database
+     my $count = $erdb->GetCount(['Genome'], 'Genome(genus-species) LIKE ?', ['homo %']);
+ would return the number of genomes for the genus I<homo>. It is conceivable, however,
+ to use it to return records based on a join. For example,
+     my $count = $erdb->GetCount(['HasFeature', 'Genome'], 'Genome(genus-species) LIKE ?',
+                                 ['homo %']);
+ would return the number of features for genomes in the genus I<homo>. Note that
+ only the rows from the first table are counted. If the above command were
+     my $count = $erdb->GetCount(['Genome', 'Feature'], 'Genome(genus-species) LIKE ?',
+                                 ['homo %']);
+ it would return the number of genomes, not the number of genome/feature pairs.
+ =over 4
+ =item objectNames
+ Reference to a list of the objects (entities and relationships) included in the
+ query.
+ =item filter
+ A filter clause for restricting the query. The rules are the same as for the L</Get>
+ method.
+ =item params
+ Reference to a list of the parameter values to be substituted for the parameter marks
+ in the filter.
+ =item RETURN
+ Returns a count of the number of records in the first table that would satisfy
+ the query.
+ =back
+ =cut
+ sub GetCount {
+     # Get the parameters.
+     my ($self, $objectNames, $filter, $params) = @_;
+     # Insure the params argument is an array reference if the caller left it off.
+     if (! defined($params)) {
+         $params = [];
+     }
+     # Declare the return variable.
+     my $retVal;
+     # Find out if we're counting an entity or a relationship.
+     my $countedField;
+     if ($self->IsEntity($objectNames->[0])) {
+         $countedField = "id";
+     } else {
+         # For a relationship we count the to-link because it's usually more
+         # numerous. Note we're automatically converting to the SQL form
+         # of the field name (to_link vs. to-link).
+         $countedField = "to_link";
+     }
+     # Create the SQL command suffix to get the desired records.
+     my ($suffix, $mappedNameListRef, $mappedNameHashRef) = $self->_SetupSQL($objectNames,
+                                                                             $filter);
+     # Prefix it with text telling it we want a record count.
+     my $firstObject = $mappedNameListRef->[0];
+     my $command = "SELECT COUNT($firstObject.$countedField) $suffix";
+     # Prepare and execute the command.
+     my $sth = $self->_GetStatementHandle($command, $params);
+     # Get the count value.
+     ($retVal) = $sth->fetchrow_array();
+     # Check for a problem.
+     if (! defined($retVal)) {
+         if ($sth->err) {
+             # Here we had an SQL error.
+             Confess("Error retrieving row count: " . $sth->errstr());
+         } else {
+             # Here we have no result.
+             Confess("No result attempting to retrieve row count.");
+         }
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 ComputeObjectSentence
+     my $sentence = $erdb->ComputeObjectSentence($objectName);
+ Check an object name, and if it is a relationship convert it to a relationship sentence.
+ =over 4
+ =item objectName
+ Name of the entity or relationship.
+ =item RETURN
+ Returns a string containing the entity name or a relationship sentence.
+ =back
+ =cut
+ sub ComputeObjectSentence {
+     # Get the parameters.
+     my ($self, $objectName) = @_;
+     # Set the default return value.
+     my $retVal = $objectName;
+     # Look for the object as a relationship.
+     my $relTable = $self->{_metaData}->{Relationships};
+     if (exists $relTable->{$objectName}) {
+         # Get the relationship sentence.
+         $retVal = _ComputeRelationshipSentence($objectName, $relTable->{$objectName});
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 DumpRelations
+     $erdb->DumpRelations($outputDirectory);
+ Write the contents of all the relations to tab-delimited files in the specified directory.
+ Each file will have the same name as the relation dumped, with an extension of DTX.
+ =over 4
+ =item outputDirectory
+ Name of the directory into which the relation files should be dumped.
+ =back
+ =cut
+ sub DumpRelations {
+     # Get the parameters.
+     my ($self, $outputDirectory) = @_;
+     # Now we need to run through all the relations. First, we loop through the entities.
+     my $metaData = $self->{_metaData};
+     my $entities = $metaData->{Entities};
+     for my $entityName (keys %{$entities}) {
+         my $entityStructure = $entities->{$entityName};
+         # Get the entity's relations.
+         my $relationList = $entityStructure->{Relations};
+         # Loop through the relations, dumping them.
+         for my $relationName (keys %{$relationList}) {
+             my $relation = $relationList->{$relationName};
+             $self->_DumpRelation($outputDirectory, $relationName, $relation);
+         }
+     }
+     # Next, we loop through the relationships.
+     my $relationships = $metaData->{Relationships};
+     for my $relationshipName (keys %{$relationships}) {
+         my $relationshipStructure = $relationships->{$relationshipName};
+         # Dump this relationship's relation.
+         $self->_DumpRelation($outputDirectory, $relationshipName, $relationshipStructure->{Relations}->{$relationshipName});
+     }
+ }
+ =head3 InsertValue
+     $erdb->InsertValue($entityID, $fieldName, $value);
+ This method will insert a new value into the database. The value must be one
+ associated with a secondary relation, since primary values cannot be inserted:
+ they occur exactly once. Secondary values, on the other hand, can be missing
+ or multiply-occurring.
+ =over 4
+ =item entityID
+ ID of the object that is to receive the new value.
+ =item fieldName
+ Field name for the new value-- this includes the entity name, since
+ field names are of the format I<objectName>C<(>I<fieldName>C<)>.
+ =item value
+ New value to be put in the field.
+ =back
+ =cut
+ sub InsertValue {
+     # Get the parameters.
+     my ($self, $entityID, $fieldName, $value) = @_;
+     # Parse the entity name and the real field name.
+     if ($fieldName =~ /^([^(]+)\(([^)]+)\)/) {
+         my $entityName = $1;
+         my $fieldTitle = $2;
+         # Get its descriptor.
+         if (!$self->IsEntity($entityName)) {
+             Confess("$entityName is not a valid entity.");
+         } else {
+             my $entityData = $self->{_metaData}->{Entities}->{$entityName};
+             # Find the relation containing this field.
+             my $fieldHash = $entityData->{Fields};
+             if (! exists $fieldHash->{$fieldTitle}) {
+                 Confess("$fieldTitle not found in $entityName.");
+             } else {
+                 my $relation = $fieldHash->{$fieldTitle}->{relation};
+                 if ($relation eq $entityName) {
+                     Confess("Cannot do InsertValue on primary field $fieldTitle of $entityName.");
+                 } else {
+                     # Now we can create an INSERT statement.
+                     my $dbh = $self->{_dbh};
+                     my $fixedName = _FixName($fieldTitle);
+                     my $statement = "INSERT INTO $relation (id, $fixedName) VALUES(?, ?)";
+                     # Execute the command.
+                     $dbh->SQL($statement, 0, $entityID, $value);
+                 }
+             }
+         }
+     } else {
+         Confess("$fieldName is not a valid field name.");
+     }
+ }
+ =head3 InsertObject
+     $erdb->InsertObject($objectType, \%fieldHash);
+ Insert an object into the database. The object is defined by a type name and then a hash
+ of field names to values. Field values in the primary relation are represented by scalars.
+ (Note that for relationships, the primary relation is the B<only> relation.)
+ Field values for the other relations comprising the entity are always list references. For
+ example, the following line inserts an inactive PEG feature named C<fig|188.1.peg.1> with aliases
+ C<ZP_00210270.1> and C<gi|46206278>.
+     $erdb->InsertObject('Feature', { id => 'fig|188.1.peg.1', active => 0, feature-type => 'peg', alias => ['ZP_00210270.1', 'gi|46206278']});
+ The next statement inserts a C<HasProperty> relationship between feature C<fig|158879.1.peg.1> and
+ property C<4> with an evidence URL of C<http://seedu.uchicago.edu/query.cgi?article_id=142>.
+     $erdb->InsertObject('HasProperty', { 'from-link' => 'fig|158879.1.peg.1', 'to-link' => 4, evidence => 'http://seedu.uchicago.edu/query.cgi?article_id=142'});
+ =over 4
+ =item newObjectType
+ Type name of the object to insert.
+ =item fieldHash
+ Hash of field names to values.
+ =back
+ =cut
+ sub InsertObject {
+     # Get the parameters.
+     my ($self, $newObjectType, $fieldHash) = @_;
+     # Denote that so far we appear successful.
+     my $retVal = 1;
+     # Get the database handle.
+     my $dbh = $self->{_dbh};
+     # Get the relation list.
+     my $relationTable = $self->_GetRelationTable($newObjectType);
+     # Loop through the relations. We'll build insert statements for each one. If a relation is
+     # secondary, we may end up generating multiple insert statements. If an error occurs, we
+     # stop the loop.
+     my @relationList = keys %{$relationTable};
+     for (my $i = 0; $retVal && $i <= $#relationList; $i++) {
+         my $relationName = $relationList[$i];
+         my $relationDefinition = $relationTable->{$relationName};
+         # Get the relation's fields. For each field we will collect a value in the corresponding
+         # position of the @valueList array. If one of the fields is missing, we will add it to the
+         # @missing list.
+         my @fieldList = @{$relationDefinition->{Fields}};
+         my @fieldNameList = ();
+         my @valueList = ();
+         my @missing = ();
+         my $recordCount = 1;
+         for my $fieldDescriptor (@fieldList) {
+             # Get the field name and save it. Note we need to fix it up so the hyphens
+             # are converted to underscores.
+             my $fieldName = $fieldDescriptor->{name};
+             push @fieldNameList, _FixName($fieldName);
+             # Look for the named field in the incoming structure. Note that we are looking
+             # for the real field name, not the fixed-up one!
+             if (exists $fieldHash->{$fieldName}) {
+                 # Here we found the field. Stash it in the value list.
+                 my $value = $fieldHash->{$fieldName};
+                 push @valueList, $value;
+                 # If the value is a list, we may need to increment the record count.
+                 if (ref $value eq "ARRAY") {
+                     my $thisCount = @{$value};
+                     if ($recordCount == 1) {
+                         # Here we have our first list, so we save its count.
+                         $recordCount = $thisCount;
+                     } elsif ($recordCount != $thisCount) {
+                         # Here we have a second list, so its length has to match the
+                         # previous lists.
+                         Trace("Field $value in new $newObjectType object has an invalid list length $thisCount. Expected $recordCount.") if T(0);
+                         $retVal = 0;
+                     }
+                 }
+             } else {
+                 # Here the field is not present. Flag it as missing.
+                 push @missing, $fieldName;
+             }
+         }
+         # Only proceed if there are no missing fields.
+         if (@missing > 0) {
+             Trace("Relation $relationName for $newObjectType skipped due to missing fields: " .
+                 join(' ', @missing)) if T(1);
+         } else {
+             # Build the INSERT statement.
+             my $statement = "INSERT INTO $relationName (" . join (', ', @fieldNameList) .
+                 ") VALUES (";
+             # Create a marker list of the proper size and put it in the statement.
+             my @markers = ();
+             while (@markers < @fieldNameList) { push @markers, '?'; }
+             $statement .= join(', ', @markers) . ")";
+             # We have the insert statement, so prepare it.
+             my $sth = $dbh->prepare_command($statement);
+             Trace("Insert statement prepared: $statement") if T(3);
+             # Now we loop through the values. If a value is scalar, we use it unmodified. If it's
+             # a list, we use the current element. The values are stored in the @parameterList array.
+             my $done = 0;
+             for (my $i = 0; $i < $recordCount; $i++) {
+                 # Clear the parameter list array.
+                 my @parameterList = ();
+                 # Loop through the values.
+                 for my $value (@valueList) {
+                     # Check to see if this is a scalar value.
+                     if (ref $value eq "ARRAY") {
+                         # Here we have a list value. Pull the current entry.
+                         push @parameterList, $value->[$i];
+                     } else {
+                         # Here we have a scalar value. Use it unmodified.
+                         push @parameterList, $value;
+                     }
+                 }
+                 # Execute the INSERT statement with the specified parameter list.
+                 $retVal = $sth->execute(@parameterList);
+                 if (!$retVal) {
+                     my $errorString = $sth->errstr();
+                     Confess("Error inserting into $relationName: $errorString");
+                 } else {
+                     Trace("Insert successful using $parameterList[0].") if T(3);
+                 }
+             }
+         }
+     }
+     # Return a 1 for backward compatability.
+     return 1;
+ }
+ =head3 UpdateEntity
+     $erdb->UpdateEntity($entityName, $id, \%fields);
+ Update the values of an entity. This is an unprotected update, so it should only be
+ done if the database resides on a database server.
+ =over 4
+ =item entityName
+ Name of the entity to update. (This is the entity type.)
+ =item id
+ ID of the entity to update. If no entity exists with this ID, an error will be thrown.
+ =item fields
+ Reference to a hash mapping field names to their new values. All of the fields named
+ must be in the entity's primary relation, and they cannot any of them be the ID field.
+ =back
+ =cut
+ sub UpdateEntity {
+     # Get the parameters.
+     my ($self, $entityName, $id, $fields) = @_;
+     # Get a list of the field names being updated.
+     my @fieldList = keys %{$fields};
+     # Verify that the fields exist.
+     my $checker = $self->GetFieldTable($entityName);
+     for my $field (@fieldList) {
+         if ($field eq 'id') {
+             Confess("Cannot update the ID field for entity $entityName.");
+         } elsif ($checker->{$field}->{relation} ne $entityName) {
+             Confess("Cannot find $field in primary relation of $entityName.");
+         }
+     }
+     # Build the SQL statement.
+     my @sets = ();
+     my @valueList = ();
+     for my $field (@fieldList) {
+         push @sets, _FixName($field) . " = ?";
+         push @valueList, $fields->{$field};
+     }
+     my $command = "UPDATE $entityName SET " . join(", ", @sets) . " WHERE id = ?";
+     # Add the ID to the list of binding values.
+     push @valueList, $id;
+     # Call SQL to do the work.
+     my $rows = $self->{_dbh}->SQL($command, 0, @valueList);
+     # Check for errors.
+     if ($rows == 0) {
+         Confess("Entity $id of type $entityName not found.");
+     }
+ }
+ =head3 LoadTable
+     my $results = $erdb->LoadTable($fileName, $relationName, %options);
+ Load data from a tab-delimited file into a specified table, optionally re-creating the table
+ first.
+ =over 4
+ =item fileName
+ Name of the file from which the table data should be loaded.
+ =item relationName
+ Name of the relation to be loaded. This is the same as the table name.
+ =item options
+ A hash of load options.
+ =item RETURN
+ Returns a statistical object containing a list of the error messages.
+ =back
+ The permissible options are as follows.
+ =over 4
+ =item truncate
+ If TRUE, then the table will be erased before loading.
+ =item mode
+ Mode in which the load should operate, either C<low_priority> or C<concurrent>.
+ This option is only applicable to a MySQL database.
+ =item partial
+ If TRUE, then it is assumed that this is a partial load, and the table will not
+ be analyzed and compacted at the end.
+ =back
+ =cut
+ sub LoadTable {
+     # Get the parameters.
+     my ($self, $fileName, $relationName, %options) = @_;
+     # Create the statistical return object.
+     my $retVal = _GetLoadStats();
+     # Trace the fact of the load.
+     Trace("Loading table $relationName from $fileName") if T(2);
+     # Get the database handle.
+     my $dbh = $self->{_dbh};
+     # Get the input file size.
+     my $fileSize = -s $fileName;
+     # Get the relation data.
+     my $relation = $self->_FindRelation($relationName);
+     # Check the truncation flag.
+     if ($options{truncate}) {
+         Trace("Creating table $relationName") if T(2);
+         # Compute the row count estimate. We take the size of the load file,
+         # divide it by the estimated row size, and then multiply by 2 to
+         # leave extra room. We postulate a minimum row count of 1000 to
+         # prevent problems with incoming empty load files.
+         my $rowSize = $self->EstimateRowSize($relationName);
+         my $estimate = $fileSize * 8 / $rowSize;
+         if ($estimate < 1000) {
+             $estimate = 1000;
+         }
+         # Re-create the table without its index.
+         $self->CreateTable($relationName, 0, $estimate);
+         # If this is a pre-index DBMS, create the index here.
+         if ($dbh->{_preIndex}) {
+             eval {
+                 $self->CreateIndex($relationName);
+             };
+             if ($@) {
+                 $retVal->AddMessage($@);
+             }
+         }
+     }
+     # Load the table.
+     my $rv;
+     eval {
+         $rv = $dbh->load_table(file => $fileName, tbl => $relationName, style => $options{mode});
+     };
+     if (!defined $rv) {
+         $retVal->AddMessage($@) if ($@);
+         $retVal->AddMessage("Table load failed for $relationName using $fileName: " . $dbh->error_message);
+         Trace("Table load failed for $relationName.") if T(1);
+     } else {
+         # Here we successfully loaded the table.
+         $retVal->Add("tables");
+         my $size = -s $fileName;
+         Trace("$size bytes loaded into $relationName.") if T(2);
+         $retVal->Add("bytes", $size);
+         # If we're rebuilding, we need to create the table indexes.
+         if ($options{truncate}) {
+             # Indexes are created here for PostGres. For PostGres, indexes are
+             # best built at the end. For MySQL, the reverse is true.
+             if (! $dbh->{_preIndex}) {
+                 eval {
+                     $self->CreateIndex($relationName);
+                 };
+                 if ($@) {
+                     $retVal->AddMessage($@);
+                 }
+             }
+             # The full-text index (if any) is always built last, even for MySQL.
+             # First we need to see if this table has a full-text index. Only
+             # primary relations are allowed that privilege.
+             Trace("Checking for full-text index on $relationName.") if T(2);
+             if ($self->_IsPrimary($relationName)) {
+                 $self->CreateSearchIndex($relationName);
+             }
+         }
+     }
+     # Analyze the table to improve performance.
+     if (! $options{partial}) {
+         Trace("Analyzing and compacting $relationName.") if T(3);
+         $self->Analyze($relationName);
+     }
+     Trace("$relationName load completed.") if T(3);
+     # Return the statistics.
+     return $retVal;
+ }
+ =head3 Analyze
+     $erdb->Analyze($tableName);
+ Analyze and compact a table in the database. This is useful after a load
+ to improve the performance of the indexes.
+ =over 4
+ =item tableName
+ Name of the table to be analyzed and compacted.
+ =back
+ =cut
+ sub Analyze {
+     # Get the parameters.
+     my ($self, $tableName) = @_;
+     # Analyze the table.
+     $self->{_dbh}->vacuum_it($tableName);
+ }
+ =head3 TruncateTable
+     $erdb->TruncateTable($table);
+ Delete all rows from a table quickly. This uses the built-in SQL
+ C<TRUNCATE> statement, which effectively drops and re-creates a table
+ with all its settings intact.
+ =over 4
+ =item table
+ Name of the table to be cleared.
+ =back
+ =cut
+ sub TruncateTable {
+     # Get the parameters.
+     my ($self, $table) = @_;
+     # Get the database handle.
+     my $dbh = $self->{_dbh};
+     # Execute a truncation comment.
+     $dbh->SQL("TRUNCATE TABLE $table");
+ }
+ =head3 CreateSearchIndex
+     $erdb->CreateSearchIndex($objectName);
+ Check for a full-text search index on the specified entity or relationship object, and
+ if one is required, rebuild it.
+ =over 4
+ =item objectName
+ Name of the entity or relationship to be indexed.
+ =back
+ =cut
+ sub CreateSearchIndex {
+     # Get the parameters.
+     my ($self, $objectName) = @_;
+     # Get the relation's entity/relationship structure.
+     my $structure = $self->_GetStructure($objectName);
+     # Get the database handle.
+     my $dbh = $self->{_dbh};
+     Trace("Checking for search fields in $objectName.") if T(3);
+     # Check for a searchable fields list.
+     if (exists $structure->{searchFields}) {
+         # Here we know that we need to create a full-text search index.
+         # Get an SQL-formatted field name list.
+         my $fields = join(", ", _FixNames(@{$structure->{searchFields}}));
+         # Create the index. If it already exists, it will be dropped.
+         $dbh->create_index(tbl => $objectName, idx => "search_idx",
+                            flds => $fields, kind => 'fulltext');
+         Trace("Index created for $fields in $objectName.") if T(2);
+     }
+ }
+ =head3 DropRelation
+     $erdb->DropRelation($relationName);
+ Physically drop a relation from the database.
+ =over 4
+ =item relationName
+ Name of the relation to drop. If it does not exist, this method will have
+ no effect.
+ =back
+ =cut
+ sub DropRelation {
+     # Get the parameters.
+     my ($self, $relationName) = @_;
+     # Get the database handle.
+     my $dbh = $self->{_dbh};
+     # Drop the relation. The method used here has no effect if the relation
+     # does not exist.
+     Trace("Invoking DB Kernel to drop $relationName.") if T(3);
+     $dbh->drop_table(tbl => $relationName);
+ }
+ =head3 MatchSqlPattern
+     my $matched = ERDB::MatchSqlPattern($value, $pattern);
+ Determine whether or not a specified value matches an SQL pattern. An SQL
+ pattern has two wild card characters: C<%> that matches multiple characters,
+ and C<_> that matches a single character. These can be escaped using a
+ backslash (C<\>). We pull this off by converting the SQL pattern to a
+ PERL regular expression. As per SQL rules, the match is case-insensitive.
+ =over 4
+ =item value
+ Value to be matched against the pattern. Note that an undefined or empty
+ value will not match anything.
+ =item pattern
+ SQL pattern against which to match the value. An undefined or empty pattern will
+ match everything.
+ =item RETURN
+ Returns TRUE if the value and pattern match, else FALSE.
+ =back
+ =cut
+ sub MatchSqlPattern {
+     # Get the parameters.
+     my ($value, $pattern) = @_;
+     # Declare the return variable.
+     my $retVal;
+     # Insure we have a pattern.
+     if (! defined($pattern) || $pattern eq "") {
+         $retVal = 1;
+     } else {
+         # Break the pattern into pieces around the wildcard characters. Because we
+         # use parentheses in the split function's delimiter expression, we'll get
+         # list elements for the delimiters as well as the rest of the string.
+         my @pieces = split /([_%]|\\[_%])/, $pattern;
+         # Check some fast special cases.
+         if ($pattern eq '%') {
+             # A null pattern matches everything.
+             $retVal = 1;
+         } elsif (@pieces == 1) {
+             # No wildcards, so we have a literal comparison. Note we're case-insensitive.
+             $retVal = (lc($value) eq lc($pattern));
+         } elsif (@pieces == 2 && $pieces[1] eq '%') {
+             # A wildcard at the end, so we have a substring match. This is also case-insensitive.
+             $retVal = (lc(substr($value, 0, length($pieces[0]))) eq lc($pieces[0]));
+         } else {
+             # Okay, we have to do it the hard way. Convert each piece to a PERL pattern.
+             my $realPattern = "";
+             for my $piece (@pieces) {
+                 # Determine the type of piece.
+                 if ($piece eq "") {
+                     # Empty pieces are ignored.
+                 } elsif ($piece eq "%") {
+                     # Here we have a multi-character wildcard. Note that it can match
+                     # zero or more characters.
+                     $realPattern .= ".*"
+                 } elsif ($piece eq "_") {
+                     # Here we have a single-character wildcard.
+                     $realPattern .= ".";
+                 } elsif ($piece eq "\\%" || $piece eq "\\_") {
+                     # This is an escape sequence (which is a rare thing, actually).
+                     $realPattern .= substr($piece, 1, 1);
+                 } else {
+                     # Here we have raw text.
+                     $realPattern .= quotemeta($piece);
+                 }
+             }
+             # Do the match.
+             $retVal = ($value =~ /^$realPattern$/i ? 1 : 0);
+         }
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 GetEntity
+     my $entityObject = $erdb->GetEntity($entityType, $ID);
+ Return an object describing the entity instance with a specified ID.
+ =over 4
+ =item entityType
+ Entity type name.
+ =item ID
+ ID of the desired entity.
+ =item RETURN
+ Returns a B<ERDBObject> representing the desired entity instance, or an undefined value if no
+ instance is found with the specified key.
+ =back
+ =cut
+ sub GetEntity {
+     # Get the parameters.
+     my ($self, $entityType, $ID) = @_;
+     # Create a query.
+     my $query = $self->Get([$entityType], "$entityType(id) = ?", [$ID]);
+     # Get the first (and only) object.
+     my $retVal = $query->Fetch();
+     if (T(3)) {
+         if ($retVal) {
+             Trace("Entity $entityType \"$ID\" found.");
+         } else {
+             Trace("Entity $entityType \"$ID\" not found.");
+         }
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 GetChoices
+     my @values = $erdb->GetChoices($entityName, $fieldName);
+ Return a list of all the values for the specified field that are represented in the
+ specified entity.
+ Note that if the field is not indexed, then this will be a very slow operation.
+ =over 4
+ =item entityName
+ Name of an entity in the database.
+ =item fieldName
+ Name of a field belonging to the entity. This is a raw field name without
+ the standard parenthesized notation used in most calls.
+ =item RETURN
+ Returns a list of the distinct values for the specified field in the database.
+ =back
+ =cut
+ sub GetChoices {
+     # Get the parameters.
+     my ($self, $entityName, $fieldName) = @_;
+     # Declare the return variable.
+     my @retVal;
+     # Get the entity data structure.
+     my $entityData = $self->_GetStructure($entityName);
+     # Get the field.
+     my $fieldHash = $entityData->{Fields};
+     if (! exists $fieldHash->{$fieldName}) {
+         Confess("$fieldName not found in $entityName.");
+     } else {
+         # Get the name of the relation containing the field.
+         my $relation = $fieldHash->{$fieldName}->{relation};
+         # Fix up the field name.
+         my $realName = _FixName($fieldName);
+         # Get the database handle.
+         my $dbh = $self->{_dbh};
+         # Query the database.
+         my $results = $dbh->SQL("SELECT DISTINCT $realName FROM $relation");
+         # Clean the results. They are stored as a list of lists, and we just want the one list.
+         @retVal = sort map { $_->[0] } @{$results};
+     }
+     # Return the result.
+     return @retVal;
+ }
+ =head3 GetEntityValues
+     my @values = $erdb->GetEntityValues($entityType, $ID, \@fields);
+ Return a list of values from a specified entity instance. If the entity instance
+ does not exist, an empty list is returned.
+ =over 4
+ =item entityType
  Entity type name.
  =item ID
+ ID of the desired entity.
+ =item fields
+ List of field names, each of the form I<objectName>C<(>I<fieldName>C<)>.
+ =item RETURN
+ Returns a flattened list of the values of the specified fields for the specified entity.
+ =back
+ =cut
+ sub GetEntityValues {
+     # Get the parameters.
+     my ($self, $entityType, $ID, $fields) = @_;
+     # Get the specified entity.
+     my $entity = $self->GetEntity($entityType, $ID);
+     # Declare the return list.
+     my @retVal = ();
+     # If we found the entity, push the values into the return list.
+     if ($entity) {
+         push @retVal, $entity->Values($fields);
+     }
+     # Return the result.
+     return @retVal;
+ }
+ =head3 GetAll
+     my @list = $erdb->GetAll(\@objectNames, $filterClause, \@parameters, \@fields, $count);
+ Return a list of values taken from the objects returned by a query. The first three
+ parameters correspond to the parameters of the L</Get> method. The final parameter is
+ a list of the fields desired from each record found by the query. The field name
+ syntax is the standard syntax used for fields in the B<ERDB> system--
+ B<I<objectName>(I<fieldName>)>-- where I<objectName> is the name of the relevant entity
+ or relationship and I<fieldName> is the name of the field.
+ The list returned will be a list of lists. Each element of the list will contain
+ the values returned for the fields specified in the fourth parameter. If one of the
+ fields specified returns multiple values, they are flattened in with the rest. For
+ example, the following call will return a list of the features in a particular
+ spreadsheet cell, and each feature will be represented by a list containing the
+ feature ID followed by all of its essentiality determinations.
+     @query = $erdb->Get(['ContainsFeature', 'Feature'], "ContainsFeature(from-link) = ?", [$ssCellID], ['Feature(id)', 'Feature(essential)']);
+ =over 4
+ =item objectNames
+ List containing the names of the entity and relationship objects to be retrieved.
+ =item filterClause
+ WHERE/ORDER BY clause (without the WHERE) to be used to filter and sort the query. The WHERE clause can
+ be parameterized with parameter markers (C<?>). Each field used must be specified in the standard form
+ B<I<objectName>(I<fieldName>)>. Any parameters specified in the filter clause should be added to the
+ parameter list as additional parameters. The fields in a filter clause can come from primary
+ entity relations, relationship relations, or secondary entity relations; however, all of the
+ entities and relationships involved must be included in the list of object names.
+ =item parameterList
+ List of the parameters to be substituted in for the parameters marks in the filter clause.
+ =item fields
+ List of the fields to be returned in each element of the list returned.
+ =item count
+ Maximum number of records to return. If omitted or 0, all available records will be returned.
+ =item RETURN
+ Returns a list of list references. Each element of the return list contains the values for the
+ fields specified in the B<fields> parameter.
+ =back
+ =cut
+ #: Return Type @@;
+ sub GetAll {
+     # Get the parameters.
+     my ($self, $objectNames, $filterClause, $parameterList, $fields, $count) = @_;
+     # Translate the parameters from a list reference to a list. If the parameter
+     # list is a scalar we convert it into a singleton list.
+     my @parmList = ();
+     if (ref $parameterList eq "ARRAY") {
+         Trace("GetAll parm list is an array.") if T(4);
+         @parmList = @{$parameterList};
+     } else {
+         Trace("GetAll parm list is a scalar: $parameterList.") if T(4);
+         push @parmList, $parameterList;
+     }
+     # Insure the counter has a value.
+     if (!defined $count) {
+         $count = 0;
+     }
+     # Add the row limit to the filter clause.
+     if ($count > 0) {
+         $filterClause .= " LIMIT $count";
+     }
+     # Create the query.
+     my $query = $self->Get($objectNames, $filterClause, \@parmList);
+     # Set up a counter of the number of records read.
+     my $fetched = 0;
+     # Loop through the records returned, extracting the fields. Note that if the
+     # counter is non-zero, we stop when the number of records read hits the count.
+     my @retVal = ();
+     while (($count == 0 || $fetched < $count) && (my $row = $query->Fetch())) {
+         my @rowData = $row->Values($fields);
+         push @retVal, \@rowData;
+         $fetched++;
+     }
+     # Return the resulting list.
+     return @retVal;
+ }
+ =head3 Exists
+     my $found = $sprout->Exists($entityName, $entityID);
+ Return TRUE if an entity exists, else FALSE.
+ =over 4
+ =item entityName
+ Name of the entity type (e.g. C<Feature>) relevant to the existence check.
+ =item entityID
+ ID of the entity instance whose existence is to be checked.
+ =item RETURN
+ Returns TRUE if the entity instance exists, else FALSE.
+ =back
+ =cut
+ #: Return Type $;
+ sub Exists {
+     # Get the parameters.
+     my ($self, $entityName, $entityID) = @_;
+     # Check for the entity instance.
+     Trace("Checking existence of $entityName with ID=$entityID.") if T(4);
+     my $testInstance = $self->GetEntity($entityName, $entityID);
+     # Return an existence indicator.
+     my $retVal = ($testInstance ? 1 : 0);
+     return $retVal;
+ }
+ =head3 EstimateRowSize
+     my $rowSize = $erdb->EstimateRowSize($relName);
+ Estimate the row size of the specified relation. The estimated row size is computed by adding
+ up the average length for each data type.
+ =over 4
+ =item relName
+ Name of the relation whose estimated row size is desired.
+ =item RETURN
+ Returns an estimate of the row size for the specified relation.
+ =back
+ =cut
+ #: Return Type $;
+ sub EstimateRowSize {
+     # Get the parameters.
+     my ($self, $relName) = @_;
+     # Declare the return variable.
+     my $retVal = 0;
+     # Find the relation descriptor.
+     my $relation = $self->_FindRelation($relName);
+     # Get the list of fields.
+     for my $fieldData (@{$relation->{Fields}}) {
+         # Get the field type and add its length.
+         my $fieldLen = $TypeTable{$fieldData->{type}}->{avgLen};
+         $retVal += $fieldLen;
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 GetFieldTable
+     my $fieldHash = $self->GetFieldTable($objectnName);
+ Get the field structure for a specified entity or relationship.
+ =over 4
+ =item objectName
+ Name of the desired entity or relationship.
+ =item RETURN
+ The table containing the field descriptors for the specified object.
+ =back
+ =cut
+ sub GetFieldTable {
+     # Get the parameters.
+     my ($self, $objectName) = @_;
+     # Get the descriptor from the metadata.
+     my $objectData = $self->_GetStructure($objectName);
+     # Return the object's field table.
+     return $objectData->{Fields};
+ }
+ =head3 SplitKeywords
+     my @keywords = ERDB::SplitKeywords($keywordString);
+ This method returns a list of the positive keywords in the specified
+ keyword string. All of the operators will have been stripped off,
+ and if the keyword is preceded by a minus operator (C<->), it will
+ not be in the list returned. The idea here is to get a list of the
+ keywords the user wants to see. The list will be processed to remove
+ duplicates.
+ It is possible to create a string that confuses this method. For example
+     frog toad -frog
+ would return both C<frog> and C<toad>. If this is a problem we can deal
+ with it later.
+ =over 4
+ =item keywordString
+ The keyword string to be parsed.
+ =item RETURN
+ Returns a list of the words in the keyword string the user wants to
+ see.
+ =back
+ =cut
+ sub SplitKeywords {
+     # Get the parameters.
+     my ($keywordString) = @_;
+     # Make a safety copy of the string. (This helps during debugging.)
+     my $workString = $keywordString;
+     # Convert operators we don't care about to spaces.
+     $workString =~ tr/+"()<>/ /;
+     # Split the rest of the string along space boundaries. Note that we
+     # eliminate any words that are zero length or begin with a minus sign.
+     my @wordList = grep { $_ && substr($_, 0, 1) ne "-" } split /\s+/, $workString;
+     # Use a hash to remove duplicates.
+     my %words = map { $_ => 1 } @wordList;
+     # Return the result.
+     return sort keys %words;
+ }
+ =head3 ValidateFieldName
+     my $okFlag = ERDB::ValidateFieldName($fieldName);
+ Return TRUE if the specified field name is valid, else FALSE. Valid field names must
+ be hyphenated words subject to certain restrictions.
+ =over 4
+ =item fieldName
+ Field name to be validated.
+ =item RETURN
+ Returns TRUE if the field name is valid, else FALSE.
+ =back
+ =cut
+ sub ValidateFieldName {
+     # Get the parameters.
+     my ($fieldName) = @_;
+     # Declare the return variable. The field name is valid until we hear
+     # differently.
+     my $retVal = 1;
+     # Compute the maximum name length.
+     my $maxLen = $TypeTable{'name-string'}->{maxLen};
+     # Look for bad stuff in the name.
+     if ($fieldName =~ /--/) {
+         # Here we have a doubled minus sign.
+         Trace("Field name $fieldName has a doubled hyphen.") if T(1);
+         $retVal = 0;
+     } elsif ($fieldName !~ /^[A-Za-z]/) {
+         # Here the field name is missing the initial letter.
+         Trace("Field name $fieldName does not begin with a letter.") if T(1);
+         $retVal = 0;
+     } elsif (length($fieldName) > $maxLen) {
+         # Here the field name is too long.
+         Trace("Maximum field name length is $maxLen. Field name must be truncated to " . substr($fieldName,0, $maxLen) . ".");
+     } else {
+         # Strip out the minus signs. Everything remaining must be a letter,
+         # underscore, or digit.
+         my $strippedName = $fieldName;
+         $strippedName =~ s/-//g;
+         if ($strippedName !~ /^(\w|\d)+$/) {
+             Trace("Field name $fieldName contains illegal characters.") if T(1);
+             $retVal = 0;
+         }
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 ReadMetaXML
+     my $rawMetaData = ERDB::ReadDBD($fileName);
+ This method reads a raw database definition XML file and returns it.
+ Normally, the metadata used by the ERDB system has been processed and
+ modified to make it easier to load and retrieve the data; however,
+ this method can be used to get the data in its raw form.
+ =over 4
+ =item fileName
+ Name of the XML file to read.
+ =item RETURN
+ Returns a hash reference containing the raw XML data from the specified file.
+ =back
+ =cut
+ sub ReadMetaXML {
+     # Get the parameters.
+     my ($fileName) = @_;
+     # Read the XML.
+     my $retVal = XML::Simple::XMLin($fileName, %XmlOptions, %XmlInOpts);
+     Trace("XML metadata loaded from file $fileName.") if T(1);
+     # Return the result.
+     return $retVal;
+ }
+ =head3 GetEntityFieldHash
+     my $fieldHashRef = ERDB::GetEntityFieldHash($structure, $entityName);
+ Get the field hash of the named entity in the specified raw XML structure.
+ The field hash may not exist, in which case we need to create it.
+ =over 4
+ =item structure
+ Raw XML structure defininng the database. This is not the run-time XML used by
+ an ERDB object, since that has all sorts of optimizations built-in.
+ =item entityName
+ Name of the entity whose field structure is desired.
+ =item RETURN
+ Returns the field hash used to define the entity's fields.
+ =back
+ =cut
+ sub GetEntityFieldHash {
+     # Get the parameters.
+     my ($structure, $entityName) = @_;
+     # Get the entity structure.
+     my $entityData = $structure->{Entities}->{$entityName};
+     # Look for a field structure.
+     my $retVal = $entityData->{Fields};
+     # If it doesn't exist, create it.
+     if (! defined($retVal)) {
+         $entityData->{Fields} = {};
+         $retVal = $entityData->{Fields};
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 WriteMetaXML
+     ERDB::WriteMetaXML($structure, $fileName);
+ Write the metadata XML to a file. This method is the reverse of L</ReadMetaXML>, and is
+ used to update the database definition. It must be used with care, however, since it
+ will only work on a raw structure, not on the processed structure created by an ERDB
+ constructor.
+ =over 4
+ =item structure
- ID of the desired entity.
+ XML structure to be written to the file.
+ =item fileName
+ Name of the output file to which the updated XML should be stored.
+ =back
+ =cut
+ sub WriteMetaXML {
+     # Get the parameters.
+     my ($structure, $fileName) = @_;
+     # Compute the output.
+     my $fileString = XML::Simple::XMLout($structure, %XmlOptions, %XmlOutOpts);
+     # Write it to the file.
+     my $xmlOut = Open(undef, ">$fileName");
+     print $xmlOut $fileString;
+ }
+ =head3 HTMLNote
+ Convert a note or comment to HTML by replacing some bulletin-board codes with HTML. The codes
+ supported are C<[b]> for B<bold>, C<[i]> for I<italics>, and C<[p]> for a new paragraph.
+ Except for C<[p]>, all the codes are closed by slash-codes. So, for
+ example, C<[b]Feature[/b]> displays the string C<Feature> in boldface.
+     my $realHtml = ERDB::HTMLNote($dataString);
+ =over 4
+ =item dataString
+ String to convert to HTML.
  =item RETURN
- Returns a B<DBObject> representing the desired entity instance, or an undefined value if no
+ An HTML string derived from the input string.
- instance is found with the specified key.
  =back
  =cut
- sub GetEntity {
+ sub HTMLNote {
+     # Get the parameter.
+     my ($dataString) = @_;
+     # HTML-escape the text.
+     my $retVal = CGI::escapeHTML($dataString);
+     # Substitute the bulletin board codes.
+     $retVal =~ s!\[(/?[bi])\]!<$1>!g;
+     $retVal =~ s!\[p\]!</p><p>!g;
+     $retVal =~ s!\[link\s+([^\]]+)\]!<a href="$1">!g;
+     $retVal =~ s!\[/link\]!</a>!g;
+     # Return the result.
+     return $retVal;
+ }
+ =head3 WikiNote
+ Convert a note or comment to Wiki text by replacing some bulletin-board codes with HTML. The codes
+ supported are C<[b]> for B<bold>, C<[i]> for I<italics>, and C<[p]> for a