Forum OpenACS Q&A: Postgres Oracle Parser

Collapse
Posted by Jason Jones on
I was just reading Philip's responses to Slashdot questions and he mentioned that he supports the OpenACS effort, including providing a server for OpenACS use and offering the Postgres development team money to implement an Oracle SQL Parser. Does anyone know if this offer was accepted, or is the Postgres team way too busy?

Personally I think database abstraction on the part of the ACS is the way to go, but I'm sure a PostgreSQL Oracle parser would make the porting more straightforward and allow OpenACS to easily keep up with Oracle/ACS, as well as helping migration to and from Oracle -- PostgreSQL in general.

Collapse
Posted by Ben Adida on
Ah yes. This goes into the "Philip Wishful Thinking" category. There are a few problems with that. First, the Postgres team didn't accept Philip's offer, as they had better things to do than write a new parser. Second, the Oracle syntax is not necessarily the best syntax to copy from a technical standpoint.

The Postgres team is implementing outer joins with an Oracle compatibility mode, so that's good. There are bogus things in Oracle, though, like the varchar(4000) limit which forces CLOB usage. Postgres's lztext extension will give us 32K varchar storage, which is exactly the size of a textarea. Now *that* is web-savvy. In the end, it's very conceivable that Postgres will be a more web-savvy database than Oracle.

Collapse
Posted by Don Baccus on
Oracle-compatible outer joins will make porting modules over easier, for sure.  The PG team's going to support SQL92 outer joins, too.  It happens that Interbase uses the SQL92 syntax, so we'll probably end up  rewriting Oracle outer joins to use standard syntax anyway if we end up deciding to support both databases.

The PG team and external contributors have added some nice Oracle-compatible stuff, i.e. to_char() most significantly.