Re: [Gusdev-gusdev] Porting to PostgreSQL

[Jason H. <jha...@pr...>]

Hi Steve,

> I suggested that we  factor those queries out of the DbiTable.pm and 
> DbiDatabase.pm files, and place them into an Oracle.pm and 
> PostgreSQL.pm.
>
> How did you guys integrate the PostgreSQL specific queries into the 
> object layer?
>
> If you haven't done a clean refactoring we can work together on that...
>
> steve
>

It looks like I have a rough first pass refactoring for Oracle and 
PostgreSQL. It still relies on checking the dsn to decide whether to 
use the Oracle of PostgreSQL specific sql outputter I wrote. I 
implemented it as a method for DbiDatabase called getDbSpecSql. This 
method creates a new instance of an GUS::ObjRelP::Oracle or 
::PostgreSQL class. These classes have methods for giving the correct 
sql statements. This could be done for DbiTable, as well, I just don't 
have it that way right now.

So far it seems to build GUS for PostgreSQL. How many people are using 
Sybase? I ask because some of the places that call for db specific sql 
ask whether you are using Sybase and others don't. However, I haven't 
written anything for Sybase, since I don't really see too much of that 
stuff around.

What is the consensus on the case sensitivity for fetchrow_hashref? If 
it is enough to merely make sure that it is called with either 
'NAME_lc' or 'NAME_uc' that would be great by me. Anybody have any 
thoughts on this?

Jason