Re: [Treebase-devel] we may have to give up on PK auto-generation by postgres

[William P. <wil...@ya...>]

On Jan 27, 2010, at 4:48 PM, Vladimir Gapeyev wrote:

> The original version of this schema (in SVN since mid-June-2009) did not have any auto-increment declarations on PKs, as far as I can see.  

Is this the DB2 schema, or the Pg schema?  Rutger and Mark J. D. converted the DB2 schema into the Pg schema -- and maybe that's what you're looking at. All existing migrated records (i.e., where all tables have id numbers that start at one and increase by one -- no big jumps and no order-of-magnitude-bigger) were done on the DB2 schema. So it's possible that: (1) the DB2 schema had a different sequencing system; (2) when the Pg schema was first designed, we forgot to include auto-increment annotations; (3) the auto-increment was later added by Rutger, but that (4) something wasn't working right with them, and hence we switched to database-wide hibernate_sequence.  

If it is the case that the original DB2 way of doing this allows us to bypass the DAO layer, can we revert to that in our Pg deployment? (Rutger -- can you give Vladimir access to the SDSC DB2 schema for him to see how this was done?)

bp