Hi Nitai, Ok, so I will add a column, just a single bit, called 'indexed' to each table that requires cataloging.
At a certain interval, say 10 Minutes, I run a script looking for records that are not flagged as indexed. I loop through this query, first adding the record to the catalog, THEN flipping the bit in the table. (this way if it errors before adding the record, it won't get flagged). Makes total sense. Now, how about getting records out of the catalog. When someone deletes something, I just delete the record. There will be too much activity to hold onto things. So if someone deletes a redirect, they get asked to confirm they really want to, and when they click yes, I just delete the record. Now, since that record is gone, it won't have an index bit, cause it is gone, but there wouldn't be anything telling the catalog to remove it, and I can't remove it remotely. Would it be best to create a table, just a simple table called something like "deleteFromIndex" where it simply lists the primary key unique ID of the record that was deleted, and as part of the catalog update script, a loop is ran on this table deleting any records from the catalog that match with the deletedtable. Then once it is confirmed that the catalog deleted the record, remove that entry from the table as well. logical? -- Open BlueDragon Public Mailing List http://www.openbluedragon.org/ http://twitter.com/OpenBlueDragon mailing list - http://groups.google.com/group/openbd?hl=en !! save a network - please trim replies before posting !!
