Re: Array with a lower bound>1, PostgreSQL 8.3.6

On Tue, Feb 24, 2009 at 09:15:40PM +0100, Federico Di Gregorio wrote:

> > Intuitively I'd expect result columns to behave in a similar
> > way, with the result array looking like:
> > 
> > [None  for x in xrange(99) ].append(1)
> 
> I don't agree on that. An array starting at 0 with 99 NULL values (None
> is almost always mapped to NULL) is completely different from an array
> of 1 value starting at 100.
That's why the dicts-of-dicts idea where the array indices
are the dict keys -- because they are
considered-to-be-sequential labels rather than true indices.

> I don't have a clear idea about how to map
> the array but surely I want the _length_ be the same in both postgresql
> and Python.
The dict would do that, wouldn't it ?

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346