Diff of /Sprout/ERDB.pm

-revision 1.33, Sat Jan 28 09:36:47 2006 UTC
+revision 1.92, Mon Jun 11 18:51:23 2007 UTC
 Line 6
      use Data::Dumper;
      use XML::Simple;
      use DBQuery;
-     use DBObject;
+     use ERDBObject;
      use Stats;
      use Time::HiRes qw(gettimeofday);
-     use FIG;
+     use Digest::MD5 qw(md5_base64);
+     use CGI;
  =head1 Entity-Relationship Database Package
-Line 58
+Line 59
  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 90
+Line 91
 -bit signed integer
+ =item counter
+-bit unsigned integer
  =item date
 -bit unsigned integer, representing a PERL date/time value
-Line 109
+Line 114
  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 125
+Line 134
  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 170
+Line 190
  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 189
+Line 211
  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 222
+Line 257
  =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 266
+Line 303
  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 300
+Line 337
  # 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. "avgLen" is the average byte length for estimating
+ # record sizes. "sort" is the key modifier for the sort command, "notes" is a type description,
- # record sizes.
+ # and "indexMod", if non-zero, is the number of characters to use when the field is specified in an
- my %TypeTable = ( char =>    { sqlType => 'CHAR(1)',            maxLen => 1,            avgLen =>   1, dataGen => "StringGen('A')" },
+ # index
-                   int =>     { sqlType => 'INTEGER',            maxLen => 20,           avgLen =>   4, dataGen => "IntGen(0, 99999999)" },
+ my %TypeTable = ( char =>    { sqlType => 'CHAR(1)',            maxLen => 1,            avgLen =>   1, sort => "",
-                   string =>  { sqlType => 'VARCHAR(255)',       maxLen => 255,          avgLen => 100, dataGen => "StringGen(IntGen(10,250))" },
+                                indexMod =>   0, notes => "single ASCII character"},
-                   text =>    { sqlType => 'TEXT',               maxLen => 1000000000,   avgLen => 500, dataGen => "StringGen(IntGen(80,1000))" },
+                   int =>     { sqlType => 'INTEGER',            maxLen => 20,           avgLen =>   4, sort => "n",
-                   date =>    { sqlType => 'BIGINT',             maxLen => 80,           avgLen =>   8, dataGen => "DateGen(-7, 7, IntGen(0,1400))" },
+                                indexMod =>   0, notes => "signed 32-bit integer"},
-                   float =>   { sqlType => 'DOUBLE PRECISION',   maxLen => 40,           avgLen =>   8, dataGen => "FloatGen(0.0, 100.0)" },
+                   counter => { sqlType => 'INTEGER UNSIGNED',   maxLen => 20,           avgLen =>   4, sort => "n",
-                   boolean => { sqlType => 'SMALLINT',           maxLen => 1,            avgLen =>   1, dataGen => "IntGen(0, 1)" },
+                                indexMod =>   0, notes => "unsigned 32-bit integer"},
+                   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,           avgLen =>  10, 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,           avgLen =>  40, 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,          avgLen =>  40, 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 324
+Line 383
                     'MM' => 'many-to-many'
                   );
- # Table for interpreting string patterns.
+ # Options for XML input and output.
+ my %XmlOptions = (GroupTags =>  { Relationships => 'Relationship',
+                                   Entities => 'Entity',
+                                   Fields => 'Field',
+                                   Indexes => 'Index',
+                                   IndexFields => 'IndexField'
+                                 },
+                   KeyAttr =>    { Relationship => 'name',
+                                   Entity => 'name',
+                                   Field => 'name'
+                                 },
+                   SuppressEmpty => 1,
+                  );
- my %PictureTable = ( 'A' => "abcdefghijklmnopqrstuvwxyz",
+ my %XmlInOpts  = (
-                      '9' => "0123456789",
+                   ForceArray => ['Field', 'Index', 'IndexField', 'Relationship', 'Entity'],
-                      'X' => "abcdefghijklmnopqrstuvwxyz0123456789",
+                   ForceContent => 1,
-                      'V' => "aeiou",
+                   NormalizeSpace => 2,
-                      'K' => "bcdfghjklmnoprstvwxyz"
                     );
+ my %XmlOutOpts = (
+                   RootName => 'Database',
+                   XMLDecl => 1,
+                  );
  =head2 Public Methods
-Line 402
+Line 478
      # 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
+ C<< 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 446
+Line 565
                  # 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 470
+Line 590
          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 480
+Line 600
          # 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 513
+Line 633
          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 $retVal;
-     close HTMLOUT;
-     Trace("Built MetaData web page.") if T(3);
  }
  =head3 DumpMetaData
-Line 539
+Line 657
      return Data::Dumper::Dumper($self->{_metaData});
  }
+ =head3 CreatePPO
+ C<< 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
+ C<< 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<< $erdb->CreateTables(); >>
-Line 558
+Line 897
      # Loop through the relations.
      for my $relationName (@relNames) {
          # Create a table for this relation.
-         $self->CreateTable($relationName);
+         $self->CreateTable($relationName, 1);
          Trace("Relation $relationName created.") if T(2);
      }
  }
-Line 626
+Line 965
      Trace("Creating table $relationName: $fieldThing") if T(2);
      $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);
      }
-Line 677
+Line 1017
              my $oldString = $fieldList->[$i];
              if (length($oldString) > $maxLen) {
                  # Here it's too big, so we truncate it.
-                 Trace("Truncating field $i in relation $relName to $maxLen characters from \"$oldString\".") if T(1);
+                 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++;
              }
-Line 687
+Line 1027
      return $retVal;
  }
+ =head3 DigestFields
+ C<< $erdb->DigestFields($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
+ C<< 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<< $erdb->CreateIndex($relationName); >>
-Line 710
+Line 1123
      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.
          my $rv = $dbh->create_index(idx => $indexName, tbl => $relationName,
-                                     flds => $flds, unique => $unique);
+                                     flds => $flds, kind => $unique);
          if ($rv) {
              Trace("Index created: $indexName for $relationName ($flds)") if T(1);
          } else {
-Line 725
+Line 1156
      }
  }
- =head3 LoadTables
+ =head3 GetSecondaryFields
- C<< my $stats = $erdb->LoadTables($directoryName, $rebuild); >>
- This method will load the database tables from a directory. The tables must already have been created
+ C<< 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.
- =item rebuild
- TRUE if the tables should be dropped and rebuilt, else FALSE. This is, unfortunately, the
+ Name of the entity whose secondary fields are desired.
- only way to erase existing data in the tables, since the TRUNCATE command is not supported
- by all of the DB engines we use.
  =item RETURN
- Returns a statistical object describing the number of records read and a list of the error messages.
+ Returns a hash mapping the field names to their field types.
  =back
  =cut
- sub LoadTables {
+ sub GetSecondaryFields {
      # Get the parameters.
-     my ($self, $directoryName, $rebuild) = @_;
+     my ($self, $entityName) = @_;
-     # Start the timer.
-     my $startTime = gettimeofday;
-     # Clean any trailing slash from the directory name.
-     $directoryName =~ s!/\\$!!;
      # Declare the return variable.
-     my $retVal = Stats->new();
+     my %retVal = ();
-     # Get the relation names.
+     # Look for the entity.
-     my @relNames = $self->GetTableNames();
+     my $table = $self->GetFieldTable($entityName);
-     for my $relationName (@relNames) {
+     # Loop through the fields, pulling out the secondaries.
-         # Try to load this relation.
+     for my $field (sort keys %{$table}) {
-         my $result = $self->_LoadRelation($directoryName, $relationName, $rebuild);
+         if ($table->{$field}->{relation} ne $entityName) {
-         # Accumulate the statistics.
+             # Here we have a secondary field.
-         $retVal->Accumulate($result);
+             $retVal{$field} = $table->{$field}->{type};
      }
-     # Add the duration of the load to the statistical object.
+     }
-     $retVal->Add('duration', gettimeofday - $startTime);
+     # Return the result.
-     # Return the accumulated statistics.
+     return %retVal;
-     return $retVal;
  }
+ =head3 GetFieldRelationName
- =head3 GetTableNames
+ C<< my $name = $erdb->GetFieldRelationName($objectName, $fieldName); >>
- C<< my @names = $erdb->GetTableNames; >>
+ Return the name of the relation containing a specified field.
- Return a list of the relations required to implement this database.
+ =over 4
- =cut
+ =item objectName
- sub GetTableNames {
+ Name of the entity or relationship containing the field.
-     # Get the parameters.
-     my ($self) = @_;
-     # Get the relation list from the metadata.
-     my $relationTable = $self->{_metaData}->{RelationTable};
-     # Return the relation names.
-     return keys %{$relationTable};
- }
- =head3 GetEntityTypes
+ =item fieldName
- C<< my @names = $erdb->GetEntityTypes; >>
+ Name of the relevant field in that entity or relationship.
- Return a list of the entity type names.
+ =item RETURN
+ Returns the name of the database relation containing the field, or C<undef> if
+ the field does not exist.
+ =back
  =cut
- sub GetEntityTypes {
+ sub GetFieldRelationName {
-     # Get the database object.
+     # Get the parameters.
-     my ($self) = @_;
+     my ($self, $objectName, $fieldName) = @_;
-     # Get the entity list from the metadata object.
+     # Declare the return variable.
-     my $entityList = $self->{_metaData}->{Entities};
+     my $retVal;
-     # Return the list of entity names in alphabetical order.
+     # Get the object field table.
-     return sort keys %{$entityList};
+     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 IsEntity
+ =head3 DeleteValue
+ C<< 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
+ C<< my $stats = $erdb->LoadTables($directoryName, $rebuild); >>
+ This method will load the database tables from a directory. The tables must already have been created
+ in the database. (This can be done by calling L</CreateTables>.) The caller passes in a directory name;
+ all of the relations to be loaded must have a file in the directory with the same name as the relation
+ (optionally with a suffix of C<.dtx>). Each file must be a tab-delimited table of field values. Each
+ line of the file will be loaded as a row of the target relation table. The field values should be in
+ the same order as the fields in the relation tables generated by L</ShowMetaData>. The old data is
+ erased before the new data is loaded in.
+ A certain amount of translation automatically takes place. Ctrl-M characters are deleted, and
+ tab and new-line characters inside a field are escaped as C<\t> and C<\n>, respectively. Dates must
+ be entered as a Unix timestamp, that is, as an integer number of seconds since the base epoch.
+ =over 4
+ =item directoryName
+ Name of the directory containing the relation files to be loaded.
+ =item rebuild
+ TRUE if the tables should be dropped and rebuilt, else FALSE. This is, unfortunately, the
+ only way to erase existing data in the tables, since the TRUNCATE command is not supported
+ by all of the DB engines we use.
+ =item RETURN
+ Returns a statistical object describing the number of records read and a list of the error messages.
+ =back
+ =cut
+ sub LoadTables {
+     # Get the parameters.
+     my ($self, $directoryName, $rebuild) = @_;
+     # Start the timer.
+     my $startTime = gettimeofday;
+     # Clean any trailing slash from the directory name.
+     $directoryName =~ s!/\\$!!;
+     # Declare the return variable.
+     my $retVal = Stats->new();
+     # Get the relation names.
+     my @relNames = $self->GetTableNames();
+     for my $relationName (@relNames) {
+         # Try to load this relation.
+         my $result = $self->_LoadRelation($directoryName, $relationName, $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 = $erdb->GetTableNames; >>
+ Return a list of the relations required to implement this database.
+ =cut
+ sub GetTableNames {
+     # Get the parameters.
+     my ($self) = @_;
+     # Get the relation list from the metadata.
+     my $relationTable = $self->{_metaData}->{RelationTable};
+     # Return the relation names.
+     return keys %{$relationTable};
+ }
+ =head3 GetEntityTypes
+ C<< my @names = $erdb->GetEntityTypes; >>
+ Return a list of the entity type names.
+ =cut
+ sub GetEntityTypes {
+     # Get the database object.
+     my ($self) = @_;
+     # Get the entity list from the metadata object.
+     my $entityList = $self->{_metaData}->{Entities};
+     # Return the list of entity names in alphabetical order.
+     return sort keys %{$entityList};
+ }
+ =head3 GetDataTypes
+ C<< 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
  C<< my $flag = $erdb->IsEntity($entityName); >>
-Line 848
+Line 1463
  =head3 Get
- C<< my $query = $erdb->Get(\@objectNames, $filterClause, $param1, $param2, ..., $paramN); >>
+ C<< 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 856
+Line 1471
  following call requests all B<Genome> objects for the genus specified in the variable
  $genus.
- C<< $query = $erdb->Get(['Genome'], "Genome(genus) = ?", $genus); >>
+ C<< $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
-Line 873
+Line 1488
  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 = $erdb->Get(['Genome', 'ComesFrom', 'Source'], "Genome(genus) = ?", $genus); >>
+ C<< $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 913
+Line 1533
  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
+ C<< "Genome(genus) = ? ORDER BY Genome(species) LIMIT 10" >>
- Parameter values to be substituted into the filter clause.
+ 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
+ C<< "LIMIT 10" >>
+ =item params
+ Reference to a list of parameter values to be substituted into the filter clause.
  =item RETURN
-Line 927
+Line 1559
  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 $command = "SELECT DISTINCT " . join(".*, ", @{$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 / LIMIT thing. The important part
-         # here is we want the filter clause to be empty if there's no WHERE filter.
-         # We'll put the ORDER BY / LIMIT clauses in the following variable.
-         my $orderClause = "";
-         # Locate the ORDER BY or LIMIT verbs (if any). We use a non-greedy
-         # operator so that we find the first occurrence of either verb.
-         if ($filterString =~ m/^(.*?)\s*(ORDER BY|LIMIT)/g) {
-             # Here we have an ORDER BY or LIMIT verb. Split it off of the filter string.
-             my $pos = pos $filterString;
-             $orderClause = $2 . substr($filterString, $pos);
-             $filterString = $1;
-         }
-         # Add the filter and the join clauses (if any) to the SELECT command.
-         if ($filterString) {
-             push @joinWhere, "($filterString)";
-         }
-         if (@joinWhere) {
-             $command .= " WHERE " . join(' AND ', @joinWhere);
-         }
-         # Add the sort or limit clause (if any) to the SELECT command.
-         if ($orderClause) {
-             $command .= " $orderClause";
-         }
      }
-     Trace("SQL query: $command") if T(SQL => 4);
-     Trace("PARMS: '" . (join "', '", @params) . "'") if (T(SQL => 4) && (@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 Delete
- C<< my $stats = $erdb->Delete($entityName, $objectID); >>
- Delete an entity instance from the database. The instance is deleted along with all entity and
+ =head3 Search
- relationship instances dependent on it. The idea of dependence here is recursive. An object is
- always dependent on itself. An object is dependent if it is a 1-to-many or many-to-many
+ C<< my $query = $erdb->Search($searchExpression, $idx, \@objectNames, $filterClause, \@params); >>
- relationship connected to a dependent entity or the "to" entity connected to a 1-to-many
- dependent relationship.
+ Perform a full text search with filtering. The search will be against a specified object
+ in the object name list. That object will get an extra field containing the search
+ relevance. Note that except for the search expression, the parameters of this method are
+ the same as those for L</Get> and follow the same rules.
  =over 4
- =item entityName
+ =item searchExpression
- Name of the entity type for the instance being deleted.
+ Boolean search expression for the text fields of the target object. The default mode for
+ a Boolean search expression is OR, but we want the default to be AND, so we will
+ add a C<+> operator to each word with no other operator before it.
- =item objectID
+ =item idx
- ID of the entity instance to be deleted. If the ID contains a wild card character (C<%>),
+ Index in the I<$objectNames> list of the table to be searched in full-text mode.
- then it is presumed to by a LIKE pattern.
+ =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.
- =item testFlag
+ =item params
- If TRUE, the delete statements will be traced without being executed.
+ Reference to a list of parameter values to be substituted into the filter clause.
  =item RETURN
- Returns a statistics object indicating how many records of each particular table were
+ Returns a query object for the specified search.
- deleted.
  =back
  =cut
- #: Return Type $%;
- sub Delete {
+ sub Search {
      # Get the parameters.
-     my ($self, $entityName, $objectID, $testFlag) = @_;
+     my ($self, $searchExpression, $idx, $objectNames, $filterClause, $params) = @_;
      # Declare the return variable.
-     my $retVal = Stats->new();
+     my $retVal;
-     # Get the DBKernel object.
+     # Create a safety copy of the parameter list. Note we have to be careful to insure
+     # a parameter list exists before we copy it.
+     my @myParams = ();
+     if (defined $params) {
+         @myParams = @{$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 DISTINCT $matchClause, " . join(".*, ", @{$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 $retVal;
+ }
+ =head3 GetFlat
+ C<< my @list = $erdb->GetFlat(\@objectNames, $filterClause, \@parameterList, $field); >>
+ This is a variation of L</GetAll> that asks for only a single field per record and
+ returns a single flattened list.
+ =over 4
+ =item objectNames
+ List containing the names of the entity and relationship objects to be retrieved.
+ =item filterClause
+ WHERE/ORDER BY clause (without the WHERE) to be used to filter and sort the query. The WHERE clause can
+ be parameterized with parameter markers (C<?>). Each field used must be specified in the standard form
+ B<I<objectName>(I<fieldName>)>. Any parameters specified in the filter clause should be added to the
+ parameter list as additional parameters. The fields in a filter clause can come from primary
+ entity relations, relationship relations, or secondary entity relations; however, all of the
+ entities and relationships involved must be included in the list of object names.
+ =item parameterList
+ List of the parameters to be substituted in for the parameters marks in the filter clause.
+ =item field
+ Name of the field to be used to get the elements of the list returned.
+ =item RETURN
+ Returns a list of values.
+ =back
+ =cut
+ #: Return Type @;
+ sub GetFlat {
+     # Get the parameters.
+     my ($self, $objectNames, $filterClause, $parameterList, $field) = @_;
+     # Construct the query.
+     my $query = $self->Get($objectNames, $filterClause, $parameterList);
+     # Create the result list.
+     my @retVal = ();
+     # Loop through the records, adding the field values found to the result list.
+     while (my $row = $query->Fetch()) {
+         push @retVal, $row->Value($field);
+     }
+     # Return the list created.
+     return @retVal;
+ }
+ =head3 SpecialFields
+ C<< my %specials = $erdb->SpecialFields($entityName); >>
+ Return a hash mapping special fields in the specified entity to the value of their
+ 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 entityName
+ 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 SpecialFields {
+     # Get the parameters.
+     my ($self, $entityName) = @_;
+     # Declare the return variable.
+     my %retVal = ();
+     # Find the entity's data structure.
+     my $entityData = $self->{_metaData}->{Entities}->{$entityName};
+     # Loop through its fields, adding each special field to the return hash.
+     my $fieldHash = $entityData->{Fields};
+     for my $fieldName (keys %{$fieldHash}) {
+         my $fieldData = $fieldHash->{$fieldName};
+         if (exists $fieldData->{special}) {
+             $retVal{$fieldName} = $fieldData->{special};
+         }
+     }
+     # Return the result.
+     return %retVal;
+ }
+ =head3 Delete
+ C<< my $stats = $erdb->Delete($entityName, $objectID, %options); >>
+ Delete an entity instance from the database. The instance is deleted along with all entity and
+ relationship instances dependent on it. The definition of I<dependence> is recursive.
+ 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.
+ =over 4
+ =item entityName
+ Name of the entity type for the instance being deleted.
+ =item objectID
+ 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 options
+ A hash detailing the options for this delete operation.
+ =item RETURN
+ Returns a statistics object indicating how many records of each particular table were
+ deleted.
+ =back
+ The permissible options for this method are as follows.
+ =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, $entityName, $objectID, %options) = @_;
+     # Declare the return variable.
+     my $retVal = Stats->new();
+     # Get the DBKernel object.
      my $db = $self->{_dbh};
      # We're going to generate all the paths branching out from the starting entity. One of
      # the things we have to be careful about is preventing loops. We'll use a hash to
-Line 1104
+Line 1840
      # FROM-relationships and entities.
      my @fromPathList = ();
      my @toPathList = ();
-     # This final hash is used to remember what work still needs to be done. We push paths
+     # This final list is used to remember what work still needs to be done. We push paths
      # onto the list, then pop them off to extend the paths. We prime it with the starting
      # point. Note that we will work hard to insure that the last item on a path in the
-     # TODO list is always an entity.
+     # to-do list is always an entity.
      my @todoList = ([$entityName]);
      while (@todoList) {
          # Get the current path.
-Line 1115
+Line 1851
          # Copy it into a list.
          my @stackedPath = @{$current};
          # Pull off the last item on the path. It will always be an entity.
-         my $entityName = pop @stackedPath;
+         my $myEntityName = pop @stackedPath;
          # Add it to the alreadyFound list.
-         $alreadyFound{$entityName} = 1;
+         $alreadyFound{$myEntityName} = 1;
+         # Figure out if we need to delete this entity.
+         if ($myEntityName ne $entityName || ! $options{keepRoot}) {
          # Get the entity data.
-         my $entityData = $self->_GetStructure($entityName);
+             my $entityData = $self->_GetStructure($myEntityName);
-         # The first task is to loop through the entity's relation. A DELETE command will
+             # Loop through the entity's relations. A DELETE command will be needed for each of them.
-         # be needed for each of them.
          my $relations = $entityData->{Relations};
          for my $relation (keys %{$relations}) {
              my @augmentedList = (@stackedPath, $relation);
              push @fromPathList, \@augmentedList;
          }
+         }
          # 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 $entityName) {
+             if ($relationship->{from} eq $myEntityName) {
                  # Add the path to this relationship.
-                 my @augmentedList = (@stackedPath, $entityName, $relationshipName);
+                 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
-Line 1146
+Line 1884
                          # the current entity, so we need to stack it.
                          my @stackList = (@augmentedList, $toEntity);
                          push @fromPathList, \@stackList;
+                     } else {
+                         Trace("$toEntity ignored because it occurred previously.") if T(4);
                      }
                  }
              }
              # Now check the TO field. In this case only the relationship needs
              # deletion.
-             if ($relationship->{to} eq $entityName) {
+             if ($relationship->{to} eq $myEntityName) {
-                 my @augmentedList = (@stackedPath, $entityName, $relationshipName);
+                 my @augmentedList = (@stackedPath, $myEntityName, $relationshipName);
                  push @toPathList, \@augmentedList;
              }
          }
      }
      # 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 page from the first qualifier
+     # 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
-Line 1171
+Line 1911
      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.
+             # 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 = "DELETE FROM $target";
+             my $stmt = $db->SetUsing(@pathTables);
-             # If there's more than just the one table, we need a USING clause.
-             if (@pathTables > 1) {
-                 $stmt .= " USING " . join(", ", @pathTables[0 .. ($#pathTables - 1)]);
-             }
              # 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.
-Line 1191
+Line 1929
                  # 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.from_link";
+                 $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 $rel.$keyName = $entity2.id";
+                     $stmt .= " AND $rel.to_link = $entity2.id";
                  }
              }
              # Now we have our desired DELETE statement.
-             if ($testFlag) {
+             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 confessing
+                 # 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: $stmt") if T(3);
+                 Trace("Executing delete from $target using '$objectID'.") if T(3);
-                 my $rv = $db->SQL($stmt, 0, [$objectID]);
+                 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.
-Line 1219
+Line 1957
      return $retVal;
  }
+ =head3 Disconnect
+ C<< $erdb->Disconnect($relationshipName, $originEntityName, $originEntityID); >>
+ Disconnect an entity instance from all the objects to which it is related. This
+ will delete each relationship instance that connects to the specified entity.
+ =over 4
+ =item relationshipName
+ Name of the relationship whose instances are to be deleted.
+ =item originEntityName
+ Name of the entity that is to be disconnected.
+ =item originEntityID
+ ID of the entity that is to be disconnected.
+ =back
+ =cut
+ sub Disconnect {
+     # Get the parameters.
+     my ($self, $relationshipName, $originEntityName, $originEntityID) = @_;
+     # Get the relationship descriptor.
+     my $structure = $self->_GetStructure($relationshipName);
+     # Insure we have a relationship.
+     if (! exists $structure->{from}) {
+         Confess("$relationshipName is not a relationship in the database.");
+     } else {
+         # Get the database handle.
+         my $dbh = $self->{_dbh};
+         # We'll set this value to 1 if we find our entity.
+         my $found = 0;
+         # Loop through the ends of the relationship.
+         for my $dir ('from', 'to') {
+             if ($structure->{$dir} eq $originEntityName) {
+                 # Delete all relationship instances on this side of the entity instance.
+                 Trace("Disconnecting in $dir direction with ID \"$originEntityID\".");
+                 $dbh->SQL("DELETE FROM $relationshipName WHERE ${dir}_link = ?", 0, $originEntityID);
+                 $found = 1;
+             }
+         }
+         # Insure we found the entity on at least one end.
+         if (! $found) {
+             Confess("Entity \"$originEntityName\" does not use $relationshipName.");
+         }
+     }
+ }
+ =head3 DeleteRow
+ C<< $erdb->DeleteRow($relationshipName, $fromLink, $toLink, \%values); >>
+ Delete a row from a relationship. In most cases, only the from-link and to-link are
+ needed; however, for relationships with intersection data values can be specified
+ for the other fields using a hash.
+ =over 4
+ =item relationshipName
+ Name of the relationship from which the row is to be deleted.
+ =item fromLink
+ ID of the entity instance in the From direction.
+ =item toLink
+ ID of the entity instance in the To direction.
+ =item values
+ Reference to a hash of other values to be used for filtering the delete.
+ =back
+ =cut
+ sub DeleteRow {
+     # Get the parameters.
+     my ($self, $relationshipName, $fromLink, $toLink, $values) = @_;
+     # Create a hash of all the filter information.
+     my %filter = ('from-link' => $fromLink, 'to-link' => $toLink);
+     if (defined $values) {
+         for my $key (keys %{$values}) {
+             $filter{$key} = $values->{$key};
+         }
+     }
+     # Build an SQL statement out of the hash.
+     my @filters = ();
+     my @parms = ();
+     for my $key (keys %filter) {
+         push @filters, _FixName($key) . " = ?";
+         push @parms, $filter{$key};
+     }
+     Trace("Parms for delete row are " . join(", ", map { "\"$_\"" } @parms) . ".") if T(SQL => 4);
+     my $command = "DELETE FROM $relationshipName WHERE " .
+                   join(" AND ", @filters);
+     # Execute it.
+     my $dbh = $self->{_dbh};
+     $dbh->SQL($command, undef, @parms);
+ }
+ =head3 DeleteLike
+ C<< my $deleteCount = $erdb->DeleteLike($relName, $filter, \@parms); >>
+ Delete all the relationship rows that satisfy a particular filter condition. Unlike a normal
+ filter, only fields from the relationship itself can be used.
+ =over 4
+ =item relName
+ Name of the relationship whose records are to be deleted.
+ =item filter
+ A filter clause (L</Get>-style) for the delete query.
+ =item parms
+ Reference to a list of parameters for the filter clause.
+ =item RETURN
+ Returns a count of the number of rows deleted.
+ =back
+ =cut
+ sub DeleteLike {
+     # Get the parameters.
+     my ($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
+ C<< 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 $retVal = "sort -t\"\t\" ";
+     # Get the relation's field list.
+     my @fields = @{$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.
+                 $fieldSpec = ($i + 1) . $modifier;
+             }
+         }
+         # Add this field to the sort command.
+         $retVal .= " -k$fieldSpec";
+     }
+     # Return the result.
+     return $retVal;
+ }
  =head3 GetList
- C<< my @dbObjects = $erdb->GetList(\@objectNames, $filterClause, $param1, $param2, ..., $paramN); >>
+ C<< my @dbObjects = $erdb->GetList(\@objectNames, $filterClause, \@params); >>
  Return a list of object descriptors for the specified objects as determined by the
  specified filter clause.
-Line 1255
+Line 2265
  filter clause in general; however, odd things may happen if a sort field is from a secondary
  relation.
- =item param1, param2, ..., paramN
+ =item params
- Parameter values to be substituted into the filter clause.
+ Reference to a list of parameter values to be substituted into the filter clause.
  =item RETURN
- Returns a list of B<DBObject>s that satisfy the query conditions.
+ Returns a list of B<ERDBObject>s that satisfy the query conditions.
  =back
-Line 1269
+Line 2279
  #: Return Type @%
  sub GetList {
      # Get the parameters.
-     my ($self, $objectNames, $filterClause, @params) = @_;
+     my ($self, $objectNames, $filterClause, $params) = @_;
      # Declare the return variable.
      my @retVal = ();
      # Perform the query.
-     my $query = $self->Get($objectNames, $filterClause, @params);
+     my $query = $self->Get($objectNames, $filterClause, $params);
      # Loop through the results.
      while (my $object = $query->Fetch) {
          push @retVal, $object;
-Line 1282
+Line 2292
      return @retVal;
  }
- =head3 ComputeObjectSentence
+ =head3 GetCount
- C<< my $sentence = $erdb->ComputeObjectSentence($objectName); >>
+ C<< my $count = $erdb->GetCount(\@objectNames, $filter, \@params); >>
- Check an object name, and if it is a relationship convert it to a relationship sentence.
+ 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 = $erdb->GetCount(['Genome', 'Feature'], 'Genome(genus-species) LIKE ?',
+                                 ['homo %']);
+ it would return the number of genomes, not the number of genome/feature pairs.
  =over 4
- =item objectName
+ =item objectNames
- Name of the entity or relationship.
+ Reference to a list of the objects (entities and relationships) included in the
+ query.
+ =item filter
+ A filter clause for restricting the query. The rules are the same as for the L</Get>
+ method.
+ =item params
+ Reference to a list of the parameter values to be substituted for the parameter marks
+ in the filter.
  =item RETURN
- Returns a string containing the entity name or a relationship sentence.
+ Returns a count of the number of records in the first table that would satisfy
+ the query.
  =back
  =cut
- sub ComputeObjectSentence {
+ sub GetCount {
      # Get the parameters.
-     my ($self, $objectName) = @_;
+     my ($self, $objectNames, $filter, $params) = @_;
-     # Set the default return value.
+     # Insure the params argument is an array reference if the caller left it off.
-     my $retVal = $objectName;
+     if (! defined($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
+ C<< 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}) {
-Line 1359
+Line 2462
      }
  }
+ =head3 InsertValue
+ C<< $erdb->InsertValue($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
- C<< my $ok = $erdb->InsertObject($objectType, \%fieldHash); >>
+ C<< $erdb->InsertObject($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.
-Line 1375
+Line 2541
  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>.
- C<< $erdb->InsertObject('HasProperty', { 'from-link' => 'fig|158879.1.peg.1', 'to-link' => 4, evidence = 'http://seedu.uchicago.edu/query.cgi?article_id=142'}); >>
+ C<< $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
-Line 1387
+Line 2553
  Hash of field names to values.
- =item RETURN
- Returns 1 if successful, 0 if an error occurred.
  =back
  =cut
-Line 1489
+Line 2651
                  $retVal = $sth->execute(@parameterList);
                  if (!$retVal) {
                      my $errorString = $sth->errstr();
-                     Trace("Insert error: $errorString.") if T(0);
+                     Confess("Error inserting into $relationName: $errorString");
+                 } else {
+                     Trace("Insert successful using $parameterList[0].") if T(3);
                  }
              }
          }
      }
-     # Return the success indicator.
+     # Return a 1 for backward compatability.
-     return $retVal;
+     return 1;
+ }
+ =head3 UpdateEntity
+ C<< $erdb->UpdateEntity($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
- C<< my %results = $erdb->LoadTable($fileName, $relationName, $truncateFlag); >>
+ C<< my $results = $erdb->LoadTable($fileName, $relationName, $truncateFlag); >>
  Load data from a tab-delimited file into a specified table, optionally re-creating the table
  first.
-Line 1547
+Line 2769
          # 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 $estimate = FIG::max($fileSize * 1.5 / $rowSize, 1000);
+         my $estimate = $fileSize * 1.5 / $rowSize;
+         if ($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.
-Line 1567
+Line 2792
      };
      if (!defined $rv) {
          $retVal->AddMessage($@) if ($@);
-         $retVal->AddMessage("Table load failed for $relationName using $fileName.");
+         $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.
-Line 1575
+Line 2800
          my $size = -s $fileName;
          Trace("$size bytes loaded into $relationName.") if T(2);
          # If we're rebuilding, we need to create the table indexes.
-         if ($truncateFlag && ! $dbh->{_preIndex}) {
+         if ($truncateFlag) {
+             # 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);
              };
-Line 1583
+Line 2811
                  $retVal->AddMessage($@);
              }
          }
+             # 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.
+     Trace("Analyzing and compacting $relationName.") if T(3);
      $dbh->vacuum_it($relationName);
+     Trace("$relationName load completed.") if T(3);
      # Return the statistics.
      return $retVal;
  }
- =head3 GenerateEntity
+ =head3 CreateSearchIndex
- C<< my $fieldHash = $erdb->GenerateEntity($id, $type, \%values); >>
+ C<< $erdb->CreateSearchIndex($objectName); >>
- Generate the data for a new entity instance. This method creates a field hash suitable for
+ Check for a full-text search index on the specified entity or relationship object, and
- passing as a parameter to L</InsertObject>. The ID is specified by the callr, but the rest
+ if one is required, rebuild it.
- of the fields are generated using information in the database schema.
- 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 objectName
- ID to assign to the new entity.
+ Name of the entity or relationship to be indexed.
- =item type
+ =back
- Type name for the new entity.
+ =cut
- =item values
+ 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
+ C<< $erdb->DropRelation($relationName); >>
- Hash containing additional values that might be needed by the data generation methods (optional).
+ 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 GenerateEntity {
+ sub DropRelation {
      # Get the parameters.
-     my ($self, $id, $type, $values) = @_;
+     my ($self, $relationName) = @_;
-     # Create the return hash.
+     # Get the database handle.
-     my $this = { id => $id };
+     my $dbh = $self->{_dbh};
-     # Get the metadata structure.
+     # Drop the relation. The method used here has no effect if the relation
-     my $metadata = $self->{_metaData};
+     # does not exist.
-     # Get this entity's list of fields.
+     Trace("Invoking DB Kernel to drop $relationName.") if T(3);
-     if (!exists $metadata->{Entities}->{$type}) {
+     $dbh->drop_table(tbl => $relationName);
-         Confess("Unrecognized entity type $type in GenerateEntity.");
+ }
-     } else {
-         my $entity = $metadata->{Entities}->{$type};
+ =head3 MatchSqlPattern
-         my $fields = $entity->{Fields};
-         # Generate data from the fields.
+ C<< my $matched = ERDB::MatchSqlPattern($value, $pattern); >>
-         _GenerateFields($this, $fields, $type, $values);
+ Determine whether or not a specified value matches an SQL pattern. An SQL
+ pattern has two wild card characters: C<%> that matches multiple characters,
+ and C<_> that matches a single character. These can be escaped using a
+ backslash (C<\>). We pull this off by converting the SQL pattern to a
+ PERL regular expression. As per SQL rules, the match is case-insensitive.
+ =over 4
+ =item value
+ Value to be matched against the pattern. Note that an undefined or empty
+ value will not match anything.
+ =item pattern
+ SQL pattern against which to match the value. An undefined or empty pattern will
+ match everything.
+ =item RETURN
+ Returns TRUE if the value and pattern match, else FALSE.
+ =back
+ =cut
+ sub MatchSqlPattern {
+     # Get the parameters.
+     my ($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);
      }
-     # Return the hash created.
+             }
-     return $this;
+             # Do the match.
+             $retVal = ($value =~ /^$realPattern$/i ? 1 : 0);
+         }
+     }
+     # Return the result.
+     return $retVal;
  }
  =head3 GetEntity
-Line 1666
+Line 2994
  =item RETURN
- Returns a B<DBObject> representing the desired entity instance, or an undefined value if no
+ Returns a B<ERDBObject> representing the desired entity instance, or an undefined value if no
  instance is found with the specified key.
  =back
-Line 1677
+Line 3005
      # Get the parameters.
      my ($self, $entityType, $ID) = @_;
      # Create a query.
-     my $query = $self->Get([$entityType], "$entityType(id) = ?", $ID);
+     my $query = $self->Get([$entityType], "$entityType(id) = ?", [$ID]);
      # Get the first (and only) object.
      my $retVal = $query->Fetch();
      # Return the result.
      return $retVal;
  }
+ =head3 GetChoices
+ C<< 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
  C<< my @values = $erdb->GetEntityValues($entityType, $ID, \@fields); >>
- Return a list of values from a specified entity instance.
+ Return a list of values from a specified entity instance. If the entity instance
+ does not exist, an empty list is returned.
  =over 4
-Line 1743
+Line 3127
  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 aliases.
+ feature ID followed by all of its essentiality determinations.
- C<< $query = $erdb->Get(['ContainsFeature', 'Feature'], "ContainsFeature(from-link) = ?", [$ssCellID], ['Feature(id)', 'Feature(alias)']); >>
+ C<< @query = $erdb->Get(['ContainsFeature', 'Feature'], "ContainsFeature(from-link) = ?", [$ssCellID], ['Feature(id)', 'Feature(essential)']); >>
  =over 4
-Line 1766
+Line 3150
  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;
+         $fetched++;
+     }
+     Trace("$fetched rows returned in GetAll.") if T(SQL => 4);
+     # Return the resulting list.
+     return @retVal;
+ }
+ =head3 Exists
+ C<< 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
+ C<< 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
+ C<< 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 $objectData->{Fields};
+ }
+ =head3 SplitKeywords
+ C<< my @keywords = ERDB::SplitKeywords($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
+ C<< 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
+ C<< 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
+ C<< 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 $retVal;
+ }
+ =head3 WriteMetaXML
+ C<< ERDB::WriteMetaXML($structure, $fileName); >>
+ Write the metadata XML to a file. This method is the reverse of L</ReadMetaXML>, and is
+ used to update the database definition. It must be used with care, however, since it
+ will only work on a raw structure, not on the processed structure created by an ERDB
+ constructor.
+ =over 4
+ =item structure
+ XML structure to be written to the file.
+ =item fileName
+ Name of the output file to which the updated XML should be stored.
+ =back
+ =cut
+ sub WriteMetaXML {
+     # Get the parameters.
+     my ($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.
+ C<< my $realHtml = ERDB::HTMLNote($dataString); >>
+ =over 4
+ =item dataString
+ String to convert to HTML.
+ =item RETURN
+ An HTML string derived from the input string.
+ =back
+ =cut
+ 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 BeginTran
+ C<< $erdb->BeginTran(); >>
+ Start a database transaction.
+ =cut
+ sub BeginTran {
+     my ($self) = @_;
+     $self->{_dbh}->begin_tran();
+ }
+ =head3 CommitTran
+ C<< $erdb->CommitTran(); >>
+ Commit an active database transaction.
+ =cut
+ sub CommitTran {
+     my ($self) = @_;
+     $self->{_dbh}->commit_tran();
+ }
+ =head3 RollbackTran
+ C<< $erdb->RollbackTran(); >>
+ Roll back an active database transaction.
+ =cut
+ sub RollbackTran {
+     my ($self) = @_;
+     $self->{_dbh}->roll_tran();
+ }
+ =head3 UpdateField
+ C<< my $count = $erdb->UpdateField($objectNames, $fieldName, $oldValue, $newValue, $filter, $parms); >>
+ Update all occurrences of a specific field value to a new value. The number of rows changed will be
+ returned.
+ =over 4
+ =item fieldName
+ Name of the field in standard I<objectName>C<(>I<fieldName>C<)> format.
+ =item oldValue
+ Value to be modified. All occurrences of this value in the named field will be replaced by the
+ new value.
+ =item newValue
+ New value to be substituted for the old value when it's found.
+ =item filter
+ A standard ERDB filter clause (see L</Get>). The filter will be applied before any substitutions take place.
+ =item parms
+ Reference to a list of parameter values in the filter.
+ =item RETURN
+ Returns the number of rows modified.
+ =back
+ =cut
+ sub UpdateField {
+     # Get the parameters.
+     my ($self, $fieldName, $oldValue, $newValue, $filter, $parms) = @_;
+     # Get the object and field names from the field name parameter.
+     $fieldName =~ /^([^(]+)\(([^)]+)\)/;
+     my $objectName = $1;
+     my $realFieldName = _FixName($2);
+     # Add the old value to the filter. Note we allow the possibility that no
+     # filter was specified.
+     my $realFilter = "$fieldName = ?";
+     if ($filter) {
+         $realFilter .= " AND $filter";
+     }
+     # Format the query filter.
+     my ($suffix, $mappedNameListRef, $mappedNameHashRef) =
+         $self->_SetupSQL([$objectName], $realFilter);
+     # Create the query. Since there is only one object name, the mapped-name data is not
+     # necessary. Neither is the FROM clause.
+     $suffix =~ s/^FROM.+WHERE\s+//;
+     # Create the update statement.
+     my $command = "UPDATE $objectName SET $realFieldName = ? WHERE $suffix";
+     # Get the database handle.
+     my $dbh = $self->{_dbh};
+     # Add the old and new values to the parameter list. Note we allow the possibility that
+     # there are no user-supplied parameters.
+     my @params = ($newValue, $oldValue);
+     if (defined $parms) {
+         push @params, @{$parms};
+     }
+     # Execute the update.
+     my $retVal = $dbh->SQL($command, 0, @params);
+     # Make the funky zero a real zero.
+     if ($retVal == 0) {
+         $retVal = 0;
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head2 Data Mining Methods
+ =head3 GetUsefulCrossValues
+ C<< my @attrNames = $sprout->GetUsefulCrossValues($sourceEntity, $relationship); >>
+ Return a list of the useful attributes that would be returned by a B<Cross> call
+ from an entity of the source entity type through the specified relationship. This
+ means it will return the fields of the target entity type and the intersection data
+ fields in the relationship. Only primary table fields are returned. In other words,
+ the field names returned will be for fields where there is always one and only one
+ value.
+ =over 4
+ =item sourceEntity
+ Name of the entity from which the relationship crossing will start.
+ =item relationship
+ Name of the relationship being crossed.
+ =item RETURN
+ Returns a list of field names in Sprout field format (I<objectName>C<(>I<fieldName>C<)>.
+ =back
+ =cut
+ #: Return Type @;
+ sub GetUsefulCrossValues {
+     # Get the parameters.
+     my ($self, $sourceEntity, $relationship) = @_;
+     # Declare the return variable.
+     my @retVal = ();
+     # Determine the target entity for the relationship. This is whichever entity is not
+     # the source entity. So, if the source entity is the FROM, we'll get the name of
+     # the TO, and vice versa.
+     my $relStructure = $self->_GetStructure($relationship);
+     my $targetEntityType = ($relStructure->{from} eq $sourceEntity ? "to" : "from");
+     my $targetEntity = $relStructure->{$targetEntityType};
+     # Get the field table for the entity.
+     my $entityFields = $self->GetFieldTable($targetEntity);
+     # The field table is a hash. The hash key is the field name. The hash value is a structure.
+     # For the entity fields, the key aspect of the target structure is that the {relation} value
+     # must match the entity name.
+     my @fieldList = map { "$targetEntity($_)" } grep { $entityFields->{$_}->{relation} eq $targetEntity }
+                         keys %{$entityFields};
+     # Push the fields found onto the return variable.
+     push @retVal, sort @fieldList;
+     # Get the field table for the relationship.
+     my $relationshipFields = $self->GetFieldTable($relationship);
+     # Here we have a different rule. We want all the fields other than "from-link" and "to-link".
+     # This may end up being an empty set.
+     my @fieldList2 = map { "$relationship($_)" } grep { $_ ne "from-link" && $_ ne "to-link" }
+                         keys %{$relationshipFields};
+     # Push these onto the return list.
+     push @retVal, sort @fieldList2;
+     # Return the result.
+     return @retVal;
+ }
+ =head3 FindColumn
+ C<< my $colIndex = ERDB::FindColumn($headerLine, $columnIdentifier); >>
+ Return the location a desired column in a data mining header line. The data
+ mining header line is a tab-separated list of column names. The column
+ identifier is either the numerical index of a column or the actual column
+ name.
+ =over 4
+ =item headerLine
+ The header line from a data mining command, which consists of a tab-separated
+ list of column names.
+ =item columnIdentifier
+ Either the ordinal number of the desired column (1-based), or the name of the
+ desired column.
+ =item RETURN
+ Returns the array index (0-based) of the desired column.
+ =back
+ =cut
+ sub FindColumn {
+     # Get the parameters.
+     my ($headerLine, $columnIdentifier) = @_;
+     # Declare the return variable.
+     my $retVal;
+     # Split the header line into column names.
+     my @headers = ParseColumns($headerLine);
+     # Determine whether we have a number or a name.
+     if ($columnIdentifier =~ /^\d+$/) {
+         # Here we have a number. Subtract 1 and validate the result.
+         $retVal = $columnIdentifier - 1;
+         if ($retVal < 0 || $retVal > $#headers) {
+             Confess("Invalid column identifer \"$columnIdentifier\": value out of range.");
+         }
+     } else {
+         # Here we have a name. We need to find it in the list.
+         for (my $i = 0; $i <= $#headers && ! defined($retVal); $i++) {
+             if ($headers[$i] eq $columnIdentifier) {
+                 $retVal = $i;
+             }
+         }
+         if (! defined($retVal)) {
+             Confess("Invalid column identifier \"$columnIdentifier\": value not found.");
+         }
+     }
+     # Return the result.
+     return $retVal;
+ }
+ =head3 ParseColumns
+ C<< my @columns = ERDB::ParseColumns($line); >>
+ Convert the specified data line to a list of columns.
+ =over 4
+ =item line
+ A data mining input, consisting of a tab-separated list of columns terminated by a
+ new-line.
+ =item RETURN
+ Returns a list consisting of the column values.
+ =back
+ =cut
+ sub ParseColumns {
+     # Get the parameters.
+     my ($line) = @_;
+     # Chop off the line-end.
+     chomp $line;
+     # Split it into a list.
+     my @retVal = split(/\t/, $line);
+     # Return the result.
+     return @retVal;
+ }
+ =head2 Virtual Methods
+ =head3 _CreatePPOIndex
+ C<< my $index = ERDB::_CreatePPOIndex($indexObject); >>
+ Convert the XML for an ERDB index to the XML structure for a PPO
+ index.
+ =over 4
+ ERDB XML structure for an index.
+ =item RETURN
+ PPO XML structure for the same index.
+ =back
+ =cut
+ sub _CreatePPOIndex {
+     # Get the parameters.
+     my ($indexObject) = @_;
+     # The incoming index contains a list of the index fields in the IndexFields
+     # member. We loop through it to create the index tags.
+     my @fields = map { { label => _FixName($_->{name}) } } @{$indexObject->{IndexFields}};
+     # Wrap the fields in attribute tags.
+     my $retVal = { attribute => \@fields };
+     # Return the result.
+     return $retVal;
+ }
+ =head3 _CreatePPOField
+ C<< my $fieldXML = ERDB::_CreatePPOField($fieldName, $fieldObject); >>
+ Convert the ERDB XML structure for a field to a PPO scalar XML structure.
+ =over 4
+ =item fieldName
- List of the fields to be returned in each element of the list returned.
+ Name of the scalar field.
- =item count
+ =item fieldObject
- Maximum number of records to return. If omitted or 0, all available records will be returned.
+ ERDB XML structure describing the field.
  =item RETURN
- Returns a list of list references. Each element of the return list contains the values for the
+ Returns a PPO XML structure for the same field.
- fields specified in the B<fields> parameter.
  =back
  =cut
- #: Return Type @@;
- sub GetAll {
+ sub _CreatePPOField {
      # Get the parameters.
-     my ($self, $objectNames, $filterClause, $parameterList, $fields, $count) = @_;
+     my ($fieldName, $fieldObject) = @_;
-     # Translate the parameters from a list reference to a list. If the parameter
+     # Get the field type.
-     # list is a scalar we convert it into a singleton list.
+     my $type = $TypeTable{$fieldObject->{type}}->{sqlType};
-     my @parmList = ();
+     # Fix up the field name.
-     if (ref $parameterList eq "ARRAY") {
+     $fieldName = _FixName($fieldName);
-         @parmList = @{$parameterList};
+     # Build the scalar tag.
-     } else {
+     my $retVal = { label => $fieldName, type => $type };
-         push @parmList, $parameterList;
+     # Return the result.
-     }
+     return $retVal;
-     # 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 EstimateRowSize
+ =head3 CleanKeywords
- C<< my $rowSize = $erdb->EstimateRowSize($relName); >>
+ C<< my $cleanedString = $erdb->CleanKeywords($searchExpression); >>
- Estimate the row size of the specified relation. The estimated row size is computed by adding
+ Clean up a search expression or keyword list. This is a virtual method that may
- up the average length for each data type.
+ be overridden by the subclass. The base-class method removes extra spaces
+ and converts everything to lower case.
  =over 4
- =item relName
+ =item searchExpression
- Name of the relation whose estimated row size is desired.
+ Search expression or keyword list to clean. Note that a search expression may
+ contain boolean operators which need to be preserved. This includes leading
+ minus signs.
  =item RETURN
- Returns an estimate of the row size for the specified relation.
+ Cleaned expression or keyword list.
  =back
  =cut
- #: Return Type $;
- sub EstimateRowSize {
+ sub CleanKeywords {
      # Get the parameters.
-     my ($self, $relName) = @_;
+     my ($self, $searchExpression) = @_;
-     # Declare the return variable.
+     # Lower-case the expression and copy it into the return variable. Note that we insure we
-     my $retVal = 0;
+     # don't accidentally end up with an undefined value.
-     # Find the relation descriptor.
+     my $retVal = lc($searchExpression || "");
-     my $relation = $self->_FindRelation($relName);
+     # Remove extra spaces.
-     # Get the list of fields.
+     $retVal =~ s/\s+/ /g;
-     for my $fieldData (@{$relation->{Fields}}) {
+     $retVal =~ s/(^\s+)|(\s+$)//g;
-         # Get the field type and add its length.
-         my $fieldLen = $TypeTable{$fieldData->{type}}->{avgLen};
-         $retVal += $fieldLen;
-     }
      # Return the result.
      return $retVal;
  }
+ =head3 GetSourceObject
+ C<< my $source = $erdb->GetSourceObject($entityName); >>
+ Return the object to be used in loading special attributes of the specified entity. The
+ algorithm for loading special attributes is stored in the C<DataGen> elements of the
+ XML
  =head2 Internal Utility Methods
- =head3 GetLoadStats
+ =head3 _RelationMap
- Return a blank statistics object for use by the load methods.
+ C<< my @relationMap = _RelationMap($mappedNameHashRef, $mappedNameListRef); >>
- This is a static method.
+ Create the relation map for an SQL query. The relation map is used by B<ERDBObject>
+ to determine how to interpret the results of the query.
- =cut
+ =over 4
- sub _GetLoadStats{
+ =item mappedNameHashRef
-     return Stats->new();
- }
- =head3 GenerateFields
+ Reference to a hash that maps modified object names to real object names.
- Generate field values from a field structure and store in a specified table. The field names
+ =item mappedNameListRef
- are first sorted by pass count, certain pre-defined fields are removed from the list, and
- then we rip through them evaluation the data generation string. Fields in the primary relation
- are stored as scalars; fields in secondary relations are stored as value lists.
- This is a static method.
+ Reference to a list of modified object names in the order they appear in the
+ SELECT list.
- =over 4
+ =item RETURN
- =item this
+ Returns a list of 2-tuples. Each tuple consists of an object name as used in the
+ query followed by the actual name of that object. This enables the B<ERDBObject> to
+ determine the order of the tables in the query and which object name belongs to each
+ mapped object name. Most of the time these two values are the same; however, if a
+ relation occurs twice in the query, the relation name in the field list and WHERE
+ clause will use a mapped name (generally the actual relation name with a numeric
+ suffix) that does not match the actual relation name.
- Hash table into which the field values should be placed.
+ =back
- =item fields
+ =cut
- Field structure from which the field descriptors should be taken.
+ sub _RelationMap {
+     # Get the parameters.
+     my ($mappedNameHashRef, $mappedNameListRef) = @_;
+     # Declare the return variable.
+     my @retVal = ();
+     # Build the map.
+     for my $mappedName (@{$mappedNameListRef}) {
+         push @retVal, [$mappedName, $mappedNameHashRef->{$mappedName}];
+     }
+     # Return it.
+     return @retVal;
+ }
- =item type
- Type name of the object whose fields are being generated.
+ =head3 _SetupSQL
+ Process a list of object names and a filter clause so that they can be used to
+ build an SQL statement. This method takes in a reference to a list of object names
+ and a filter clause. It will return a corrected filter clause, a list of mapped
+ names and the mapped name hash.
+ This is an instance method.
+ =over 4
+ =item objectNames
+ Reference to a list of the object names to be included in the query.
- =item values (optional)
+ =item filterClause
- Reference to a value structure from which additional values can be taken.
+ A string containing the WHERE clause for the query (without the C<WHERE>) and also
+ optionally the C<ORDER BY> and C<LIMIT> clauses.
- =item from (optiona)
+ =item matchClause
- Reference to the source entity instance if relationship data is being generated.
+ An optional full-text search clause. If specified, it will be inserted at the
+ front of the WHERE clause. It should already be SQL-formatted; that is, the
+ field names should be in the form I<table>C<.>I<fieldName>.
- =item to (optional)
+ =item RETURN
- Reference to the target entity instance if relationship data is being generated.
+ Returns a three-element list. The first element is the SQL statement suffix, beginning
+ with the FROM clause. The second element is a reference to a list of the names to be
+ used in retrieving the fields. The third element is a hash mapping the names to the
+ objects they represent.
  =back
  =cut
- sub _GenerateFields {
+ sub _SetupSQL {
-     # Get the parameters.
+     my ($self, $objectNames, $filterClause, $matchClause) = @_;
-     my ($this, $fields, $type, $values, $from, $to) = @_;
+     # Adjust the list of object names to account for multiple occurrences of the
-     # Sort the field names by pass number.
+     # same object. We start with a hash table keyed on object name that will
-     my @fieldNames = sort { $fields->{$a}->{DataGen}->{pass} <=> $fields->{$b}->{DataGen}->{pass} } keys %{$fields};
+     # return the object suffix. The first time an object is encountered it will
-     # Loop through the field names, generating data.
+     # not be found in the hash. The next time the hash will map the object name
-     for my $name (@fieldNames) {
+     # to 2, then 3, and so forth.
-         # Only proceed if this field needs to be generated.
+     my %objectHash = ();
-         if (!exists $this->{$name}) {
+     # This list will contain the object names as they are to appear in the
-             # Get this field's data generation descriptor.
+     # FROM list.
-             my $fieldDescriptor = $fields->{$name};
+     my @fromList = ();
-             my $data = $fieldDescriptor->{DataGen};
+     # This list contains the suffixed object name for each object. It is exactly
-             # Get the code to generate the field value.
+     # parallel to the list in the $objectNames parameter.
-             my $codeString = $data->{content};
+     my @mappedNameList = ();
-             # Determine whether or not this field is in the primary relation.
+     # Finally, this hash translates from a mapped name to its original object name.
-             if ($fieldDescriptor->{relation} eq $type) {
+     my %mappedNameHash = ();
-                 # Here we have a primary relation field. Store the field value as
+     # Now we create the lists. Note that for every single name we push something into
-                 # a scalar.
+     # @fromList and @mappedNameList. This insures that those two arrays are exactly
-                 $this->{$name} = eval($codeString);
+     # parallel to $objectNames.
+     for my $objectName (@{$objectNames}) {
+         # Get the next suffix for this object.
+         my $suffix = $objectHash{$objectName};
+         if (! $suffix) {
+             # Here we are seeing the object for the first time. The object name
+             # is used as is.
+             push @mappedNameList, $objectName;
+             push @fromList, $objectName;
+             $mappedNameHash{$objectName} = $objectName;
+             # Denote the next suffix will be 2.
+             $objectHash{$objectName} = 2;
+         } else {
+             # Here we've seen the object before. We construct a new name using
+             # the suffix from the hash and update the hash.
+             my $mappedName = "$objectName$suffix";
+             $objectHash{$objectName} = $suffix + 1;
+             # The FROM list has the object name followed by the mapped name. This
+             # tells SQL it's still the same table, but we're using a different name
+             # for it to avoid confusion.
+             push @fromList, "$objectName $mappedName";
+             # The mapped-name list contains the real mapped name.
+             push @mappedNameList, $mappedName;
+             # Finally, enable us to get back from the mapped name to the object name.
+             $mappedNameHash{$mappedName} = $objectName;
+         }
+     }
+     # Begin the SELECT suffix. It starts with
+     #
+     # FROM name1, name2, ... nameN
+     #
+     my $suffix = "FROM " . join(', ', @fromList);
+     # Now for the WHERE. First, we need a place for the filter string.
+     my $filterString = "";
+     # We will also keep a list of conditions to add to the WHERE clause in order to link
+     # entities and relationships as well as primary relations to secondary ones.
+     my @joinWhere = ();
+     # Check for a filter clause.
+     if ($filterClause) {
+         # Here we have one, so we convert its field names and add it to the query. First,
+         # We create a copy of the filter string we can work with.
+         $filterString = $filterClause;
+         # Next, we sort the object names by length. This helps protect us from finding
+         # object names inside other object names when we're doing our search and replace.
+         my @sortedNames = sort { length($b) - length($a) } @mappedNameList;
+         # 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. We may
+         # need to add relations later if there is filtering on a field in a secondary
+         # relation. The secondary relations are the ones that contain multiply-
+         # occurring or optional fields.
+         my %fromNames = map { $_ => 1 } @sortedNames;
+         # 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 $mappedName (@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 $mappedName;
+             # Get the real object name for this mapped name.
+             my $objectName = $mappedNameHash{$mappedName};
+             Trace("Processing $mappedName for object $objectName.") if T(4);
+             # Get the object's field list.
+             my $fieldList = $self->GetFieldTable($objectName);
+             # Find the field references for this object.
+             while ($filterString =~ m/$mappedName\(([^)]*)\)/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 {
+                     Trace("Processing $fieldName at position $pos.") if T(4);
+                     # Get the field's relation.
+                     my $relationName = $fieldList->{$fieldName}->{relation};
+                     # Now we have a secondary relation. We need to insure it matches the
+                     # mapped name of the primary relation. First we peel off the suffix
+                     # from the mapped name.
+                     my $mappingSuffix = substr $mappedName, length($objectName);
+                     # Put the mapping suffix onto the relation name to get the
+                     # mapped relation name.
+                     my $mappedRelationName = "$relationName$mappingSuffix";
+                     # Insure the relation is in the FROM clause.
+                     if (!exists $fromNames{$mappedRelationName}) {
+                         # Add the relation to the FROM clause.
+                         if ($mappedRelationName eq $relationName) {
+                             # The name is un-mapped, so we add it without
+                             # any frills.
+                             $suffix .= ", $relationName";
+                             push @joinWhere, "$objectName.id = $relationName.id";
+                         } else {
+                             # Here we have a mapping situation.
+                             $suffix .= ", $relationName $mappedRelationName";
+                             push @joinWhere, "$mappedRelationName.id = $mappedName.id";
+                         }
+                         # Denote we have this relation available for future fields.
+                         $fromNames{$mappedRelationName} = 1;
+                     }
+                     # Form an SQL field reference from the relation name and the field name.
+                     my $sqlReference = "$mappedRelationName." . _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, because we shift off the first object before
+     # processing the rest.
+     my @mappedObjectList = @mappedNameList;
+     my $lastMappedObject = shift @mappedObjectList;
+     # Get the join table.
+     my $joinTable = $self->{_metaData}->{Joins};
+     # Loop through the object list.
+     for my $thisMappedObject (@mappedObjectList) {
+         # Look for a join using the real object names.
+         my $lastObject = $mappedNameHash{$lastMappedObject};
+         my $thisObject = $mappedNameHash{$thisMappedObject};
+         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 $lastMappedObject to $thisMappedObject.");
              } else {
-                 # Here we have a secondary relation field. Create a null list
+             # Get the join clause.
-                 # and push the desired number of field values onto it.
+             my $unMappedJoin = $joinTable->{$joinKey};
-                 my @fieldValues = ();
+             # Fix the names.
-                 my $count = IntGen(0,$data->{testCount});
+             $unMappedJoin =~ s/$lastObject/$lastMappedObject/;
-                 for (my $i = 0; $i < $count; $i++) {
+             $unMappedJoin =~ s/$thisObject/$thisMappedObject/;
-                     my $newValue = eval($codeString);
+             push @joinWhere, $unMappedJoin;
-                     push @fieldValues, $newValue;
+             # Save this object as the last object for the next iteration.
+             $lastMappedObject = $thisMappedObject;
+         }
+     }
+     # Now we need to handle the whole ORDER BY / LIMIT thing. The important part
+     # here is we want the filter clause to be empty if there's no WHERE filter.
+     # We'll put the ORDER BY / LIMIT clauses in the following variable.
+     my $orderClause = "";
+     # This is only necessary if we have a filter string in which the ORDER BY
+     # and LIMIT clauses can live.
+     if ($filterString) {
+         # Locate the ORDER BY or LIMIT verbs (if any). We use a non-greedy
+         # operator so that we find the first occurrence of either verb.
+         if ($filterString =~ m/^(.*?)\s*(ORDER BY|LIMIT)/g) {
+             # Here we have an ORDER BY or LIMIT verb. Split it off of the filter string.
+             my $pos = pos $filterString;
+             $orderClause = $2 . substr($filterString, $pos);
+             $filterString = $1;
+         }
                  }
-                 # Store the value list in the main hash.
+     # All the things that are supposed to be in the WHERE clause of the
-                 $this->{$name} = \@fieldValues;
+     # SELECT command need to be put into @joinWhere so we can string them
+     # together. We begin with the match clause. This is important,
+     # because the match clause's parameter mark must precede any parameter
+     # marks in the filter string.
+     if ($matchClause) {
+         push @joinWhere, $matchClause;
              }
+     # Add the filter string. We put it in parentheses to avoid operator
+     # precedence problems with the match clause or the joins.
+     if ($filterString) {
+         Trace("Filter string is \"$filterString\".") if T(4);
+         push @joinWhere, "($filterString)";
+     }
+     # String it all together into a big filter clause.
+     if (@joinWhere) {
+         $suffix .= " WHERE " . join(' AND ', @joinWhere);
          }
+     # Add the sort or limit clause (if any).
+     if ($orderClause) {
+         $suffix .= " $orderClause";
+     }
+     # Return the suffix, the mapped name list, and the mapped name hash.
+     return ($suffix, \@mappedNameList, \%mappedNameHash);
+ }
+ =head3 _GetStatementHandle
+ This method will prepare and execute an SQL query, returning the statement handle.
+ The main reason for doing this here is so that everybody who does SQL queries gets
+ the benefit of tracing.
+ This is an instance method.
+ =over 4
+ =item command
+ Command to prepare and execute.
+ =item params
+ Reference to a list of the values to be substituted in for the parameter marks.
+ =item RETURN
+ Returns a prepared and executed statement handle from which the caller can extract
+ results.
+ =back
+ =cut
+ sub _GetStatementHandle {
+     # Get the parameters.
+     my ($self, $command, $params) = @_;
+     # Trace the query.
+     Trace("SQL query: $command") if T(SQL => 3);
+     Trace("PARMS: '" . (join "', '", @{$params}) . "'") if (T(SQL => 4) && (@{$params} > 0));
+     # Get the database handle.
+     my $dbh = $self->{_dbh};
+     # Prepare the command.
+     my $sth = $dbh->prepare_command($command);
+     # Execute it with the parameters bound in.
+     $sth->execute(@{$params}) || Confess("SELECT error:  " . $sth->errstr());
+     # Return the statement handle.
+     return $sth;
      }
+ =head3 _GetLoadStats
+ Return a blank statistics object for use by the load methods.
+ This is a static method.
+ =cut
+ sub _GetLoadStats{
+     return Stats->new();
  }
- =head3 DumpRelation
+ =head3 _DumpRelation
- Dump the specified relation's to the specified output file in tab-delimited format.
+ Dump the specified relation to the specified output file in tab-delimited format.
  This is an instance method.
-Line 1994
+Line 4335
      close DTXOUT;
  }
- =head3 GetStructure
+ =head3 _GetStructure
  Get the data structure for a specified entity or relationship.
-Line 2033
+Line 4374
      return $retVal;
  }
- =head3 GetRelationTable
+ =head3 _GetRelationTable
  Get the list of relations for a specified entity or relationship.
-Line 2062
+Line 4405
      return $objectData->{Relations};
  }
- =head3 GetFieldTable
+ =head3 _ValidateFieldNames
- Get the field structure for a specified entity or relationship.
- This is an instance method.
- =over 4
- =item objectName
- Name of the desired entity or relationship.
- =item RETURN
- The table containing the field descriptors for the specified object.
- =back
- =cut
- sub _GetFieldTable {
-     # Get the parameters.
-     my ($self, $objectName) = @_;
-     # Get the descriptor from the metadata.
-     my $objectData = $self->_GetStructure($objectName);
-     # Return the object's field table.
-     return $objectData->{Fields};
- }
- =head3 ValidateFieldNames
  Determine whether or not the field names are valid. A description of the problems with the names
  will be written to the standard error output. If there is an error, this method will abort. This is
-Line 2118
+Line 4432
          for my $object (values %{$metadata->{$section}}) {
              # Loop through the object's fields.
              for my $fieldName (keys %{$object->{Fields}}) {
-                 # Now we make some initial validations.
+                 # If this field name is invalid, set the return value to zero
-                 if ($fieldName =~ /--/) {
+                 # so we know we encountered an error.
-                     # Here we have a doubled minus sign.
+                 if (! ValidateFieldName($fieldName)) {
-                     print STDERR "Field name $fieldName has a doubled hyphen.\n";
                      $retVal = 0;
-                 } elsif ($fieldName !~ /^[A-Za-z]/) {
-                     # Here the field name is missing the initial letter.
-                     print STDERR "Field name $fieldName does not begin with a letter.\n";
-                     $retVal = 0;
-                 } else {
-                     # Strip out the minus signs. Everything remaining must be a letter
-                     # or digit.
-                     my $strippedName = $fieldName;
-                     $strippedName =~ s/-//g;
-                     if ($strippedName !~ /^[A-Za-z0-9]+$/) {
-                         print STDERR "Field name $fieldName contains illegal characters.\n";
-                         $retVal = 0;
-                     }
                  }
              }
          }
-Line 2146
+Line 4446
      }
  }
- =head3 LoadRelation
+ =head3 _LoadRelation
  Load a relation from the data in a tab-delimited disk file. The load will only take place if a disk
  file with the same name as the relation exists in the specified directory.
-Line 2206
+Line 4506
      return $retVal;
  }
- =head3 LoadMetaData
+ =head3 _LoadMetaData
+ C<< my $metadata = ERDB::_LoadMetaData($filename); >>
  This method loads the data describing this database from an XML file into a metadata structure.
  The resulting structure is a set of nested hash tables containing all the information needed to
-Line 2231
+Line 4534
  sub _LoadMetaData {