Data Conversion from XPat to MySQL

Convert Image Class SGML to MySQL easily using the Perl script ( icsgml2icmysql.pl ) included with the release. The script produces MySQL scripts that in turn generate an indexed Data Table and an indexed Media Table in MySQL. The script handles very well collections that require a Media Table.

Converting an existing Image Class collection to MySQL involves several steps:

• Update the Collmgr record for the collection to meet DLXS 11 conventions. If you've already migrated to DLXS 11 middleware and your collections are working with XPAT, then this step has been done.
• Add administrative field mappings to the Collmgr field called field_admin_maps. See the help text in Collmgr for details.
• Also add searchable fields to the Collmgr field called dfltsrchflds.
• The fields that are to be searchable must also be configured by completing the Collmgr field called dfltsrchflds, as well as the field field_xcoll_map. Both fields are used to determine the indexes to be built.
• If the SGML predates the DLXS 10 release of Image Class, the image index directory and SGML should be regenerated. It is not a bad idea in any case.
  • Update the image index directory using imageprep script.
  • Regenerate SGML to update the tagging.
• Run the conversion script
  
$DLXSROOT/bin/i/image/icsgml2icmysql.pl collid
  .
• Run the MySQL scripts generated by icsgml2mysql.pl. See also Data Loading.
  
mysql -u dlxsadm -h yourmysqlhost -p dlxs < $DLXSROOT/prep/c/collid/ic.collid_data.sql
mysql -u dlxsadm -h yourmysqlhost -p dlxs < $DLXSROOT/prep/c/collid/ic.collid_media.sql
  
  
• Modify the collmgr record for the collection so Image Class knows to use MySQL instead of XPat
• Make minor changes to collection specific templates.

When using icsgml2mysql.pl for conversion, ic.coll.norm.sgml must have line breaks. If it does not, run idb transform and norm again to generate a file that does.

A newly made imageprep index is backward compatible with DLXS 10 SGML as long the collection only includes MrSID files and JPEG thumbnails and the filename in the metadata has the extension ".TIF". Otherwise, edit the imageprep script simply to set the initial value of $hold to '1' and then run imageprep. This forces it to generate the new index directory as "indexprep" without replacing the current "index" directory. Move the new "indexprep" directory to img/c/coll of your data prep space (assuming such space exists) and rename "indexprep" as "index". After reprocessing the metadata move the XPat and/or MySQL data into production, and rename the production "indexprep" directory as "index". Remember to change $hold back to '0' when done.

Important: Until the conversion of all collections is complete, continue to maintain the SGML/XPat version of the data. Searching of XPat and MySQL collections together defaults to XPat for all collections.