On Tue, Oct 31, 2006 at 12:55:46PM -0500, Andrew Dunstan wrote:
> To this you propose, as I understand it, to have a fourth possibility
> which would be spec compliant for comparison purposes but would label
> result set columns with the case preserved name originally used (or
> would you use the casing used in the query?).
The big issue I see with this is that it would break PQfname on the
client end, since that's case sensetive too. Most client languages are,
so you really are between a rock and a hard place.
Making PQfname case-insensetive also screws up in Tom's example.
One way to appraoch this is to consider this a setting of the collation
of the name datatype. If a case-insensetive collation is selected at
initdb time, then Tom's example would indeed fail, but that's a choice
someone made. Problem being, you'd have to export that choice to
clients to make PQfname work, and that's going to messy.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.