Another helpful thing may be to create views over your data and use sqlxml to output as xml and view in an xml visualisation tool. This also makes it easy to make data changes and then dump data back also via sqlxml or by use of a xslt sheet to build your insert statements. But at least you can see the problem visually and find make a plan from there. If unfamiliar with the db it may help to use sql workbench or whatever and rev engineer structure into uml or use d2r to export as rdf (it will find table links for you) you can then visualise in allegrograph gruff or any rdf viewer.
On Fri, Mar 12, 2010 at 11:55 AM, Javier Bezos jbezos@fundeu.es wrote:
When we have had problems it was easier to dump database to sql
statements,
drop database, recreate with media wiki setup with all correct settings
and
restore the data without drop table instructions.
Sounds sensible. Thank you.
Unfortutately, as a consequence of one of their changes, data has been partially lost (apparently only redirects and a few templates and images) and I think several tables are corrupted in the sense they contain data which shouldn't be there, so I'm even considering an export and then and import (with history, of course) for articles and the like, and only resort to the database for users (this table seems ok) and the like (just brainstrorming). <sigh>
Javier
MediaWiki-l mailing list MediaWiki-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mediawiki-l