Re: point types in "DISTINCT" queries

> On Wed, 2011-06-29 at 11:37 -0400, Jonathan S. Katz wrote:
>> Which means it *should* work, but first I would need to clean up the data and find the duplicates. I was hoping this might work:
>>
>> SELECT geocode, count(*)
>> FROM a
>> GROUP BY a.geocode
>> HAVING count(*) > 1;
>
> Maybe you could use a self-join as a workaround for now, just to clean
> up the data?
>
> SELECT geocode, other_columns from a a1, a a2 where a1.other_columns <>
> a2.other_columns and a1.geocode ~= a2.geocode;

That worked perfectly - turned out it was just two rows. And subsequently executing the exclusion constraint on "=~" also worked perfectly as expected.

The larger issue I face with now is slightly out of my control without further hacking. I'm developing an app with Django and I wrote an extension that allows me to use the point type natively in Python. I ran into the original issue while an automatically generated query was executed in the admin section. I know this could be viewed as something pertaining to Django, but the goal I had in mind was making PostgreSQL functionality more accessible in a different software layer.

I will find a workaround for the above, as I am sure I can do some application-level hacking.

Thanks for your help!

Jonathan