Hitting the limits of your tools... (...is limiting.)

I use PyTable fairly extensively (actually exclusively) when I'm working with databases. Today, however, I've come up against a limitation that I've long known to exist, but have never before had a reason to fix. Makes for a rather slow hacking day I'm afraid.

In particular, PostgreSQL has a feature which defines distinct namespaces for tables/views/indices/sequences to avoid naming conflicts among sub-systems of more complex applications. Up until a few months ago, we had nothing that actually used this feature, so I'd never had a reason to implement it.

Today, however, I started working on the database side of the new Electronic Funds Transfer (EFT) server, and the spec says that we are going to put the EFT sub-schema into a namespace to make it more composable (i.e. to make it easier to add to an existing database).

The other major feature that's been on the "should do that some day" for PyTable is views, which are far more common than namespaces, but again, I've not had a reason to support them, so they are still sitting there unimplemented. These two features define the missing functionality I want for a 1.0 release, so PyTable is still in alpha status.

Hopefully the namespace implementation will be done in a few hours, but I don't know when the view implementation will float to the top of the priority list.


Comments are closed.


Pingbacks are closed.