[postgis-users] Bogus unknown geometry error from ST_Transform

Sun Feb 19 13:19:16 PST 2012

As of 2.0 the whole question will be moot, as geometry_columns will be a
view and will accurately reflect the state of the tables... so a table with
no known srid will show up with unknown srid, and a table with known srid
will have an entry... but since it will not be possible to insert an
unknown srid geometry into a known srid table, your case won't ever arise.

P.

On Sun, Feb 19, 2012 at 1:00 PM, Aren Cambre <aren at arencambre.com> wrote:

> I hear you and am OK with reclassifying as an enhancement. However, what's
> the purpose of public.geometry_columns?
>
> If it's just ephemeral, then I accept your "do-what-I-mean-not-what-I-say"
> point.
>
> If it's supposed to be accurate metadata, then if some functionality needs
> this metadata *and* its missing from the geometry field, why not
> consult public.geometry_columns?
>
> Aren
>
>
> On Sun, Feb 19, 2012 at 2:35 PM, Paul Ramsey <pramsey at opengeo.org> wrote:
>
>> A bug is unspecified or unexpected behavior, and I'm afraid your case is
>> neither: the system is operating exactly as I would expect. You have
>> outlined what might be considered a legitimate enhancement, but it's on the
>> edge, because do-what-I-mean-not-what-I-say behavior can cause other people
>> problems from time to time.
>>
>> P.
>>
>>
>> On Sun, Feb 19, 2012 at 12:29 PM, Aren Cambre <aren at arencambre.com>wrote:
>>
>>> OK, I figured out the second error. The TIGER places dataset for Texas
>>> has both POLYGON and MULTIPOLYGON types, hence violating the *
>>> enforce_geotype_the_geom_3081* constraint.
>>>
>>> However, I think the original issue is a bug. If the geometry object has
>>> no SRID specified, as in my case, then ST_Transform should use whatever's
>>> specified in *public.geometry_columns*. I should not be getting errors
>>> about SRID of -1 if something is specified in  *public.geometry_columns*
>>> .
>>>
>>> Aren
>>>
>>> On Sun, Feb 19, 2012 at 9:13 AM, Aren Cambre <aren at arencambre.com>wrote:
>>>
>>>> Nope. BTW, there is an extraneous parenthesis in your example. Here's
>>>> what I used:
>>>> *UPDATE gis."gz_2010_48_160_00_500k"*
>>>> *SET "the_geom_3081" = ST_Transform(ST_Setsrid(the_geom, 4629),3081);*
>>>>
>>>> (BTW, I said 4326 earlier, but I should have been using 4269. Not sure
>>>> that explains the errors, however.)
>>>>
>>>> Now I get this:
>>>> *ERROR: new row for relation "gz_2010_48_160_00_500k" violates check
>>>> constraint "enforce_geotype_the_geom_3081"*
>>>>
>>>> This is just bizarre because the source and destination column are
>>>> POLYGON types.
>>>>
>>>> Here's what I get when running *ST_AsEWKT(the_geom)*:
>>>> *"POLYGON((-94.940569 29.330815,-94.93334 29.329038,-94.930971
>>>> 29.327708,-94.929679 29.32634,-94.938834 29.320412,-94.946644
>>>> 29.323065,-94.946988 29.326066,-94.94085 29.330469,-94.940569 29.330815))"
>>>> *
>>>>
>>>> I just double-checked, and there are no other rows in
>>>> public.geometry_columns describing anything in the *
>>>> gz_2010_48_160_00_500k* table besides the two rows I pasted earlier.
>>>>
>>>> Here's that constraint:
>>>> *enforce_geotype_the_geom_3081 CHECK (geometrytype(the_geom_3081) =
>>>> 'POLYGON'::text OR the_geom_3081 IS NULL);*
>>>>
>>>> Aren
>>>>
>>>> On Sun, Feb 19, 2012 at 2:10 AM, <pcreso at pcreso.com> wrote:
>>>>
>>>>> Does this work?
>>>>>
>>>>> *UPDATE gis."gz_2010_48_160_00_500k"*
>>>>> *SET "the_geom_3081" = ST_Transform((ST_Setsrid(the_geom, 4326),3081);
>>>>> *
>>>>>
>>>>> ie:: is the problem finding the srid from geometry_columns or finding
>>>>> it but ST_Transform() fails even when given the srid?
>>>>>
>>>>> Even if there is a bug, this may work for you until fixed?
>>>>>
>>>>> Brent Wood
>>>>>
>>>>> --- On *Sun, 2/19/12, René Romero Benavides <ichbinrene at gmail.com>*wrote:
>>>>>
>>>>>
>>>>> From: René Romero Benavides <ichbinrene at gmail.com>
>>>>> Subject: Re: [postgis-users] Bogus unknown geometry error from
>>>>> ST_Transform
>>>>> To: postgis-users at postgis.refractions.net
>>>>> Date: Sunday, February 19, 2012, 8:37 PM
>>>>>
>>>>>
>>>>>  El 18/02/2012 11:29 p.m., Aren Cambre escribió:
>>>>>
>>>>> ST_Transform is returning a bogus error of* Input geometry has
>>>>> unknown (-1) SRID*.
>>>>>
>>>>>  Here's the query:
>>>>>  *UPDATE gis."gz_2010_48_160_00_500k"*
>>>>> *SET "the_geom_3081" = ST_Transform(the_geom, 3081);*
>>>>>
>>>>>  I said "bogus" because the database's *public.geometry_columns*table has valid values for these two columns, and neither has SRID of -1:
>>>>>  *
>>>>> 34281;"''";"gis";"gz_2010_48_160_00_500k";"the_geom";2;4326;"POLYGON"*
>>>>> *
>>>>> 34276;"''";"gis";"gz_2010_48_160_00_500k";"the_geom_3081";2;3081;"POLYGON"
>>>>> *
>>>>>
>>>>>  What gives?
>>>>>
>>>>>  Aren
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> postgis-users mailing listpostgis-users at postgis.refractions.net <http://mc/compose?to=postgis-users@postgis.refractions.net>http://postgis.refractions.net/mailman/listinfo/postgis-users
>>>>>
>>>>>  I thought a SRID of -1 was a synonym for the default one (4326). But
>>>>> don't take my word for it, I'm beginning with postgis.
>>>>> --
>>>>>  http://sharingtechknowledge.blogspot.com/
>>>>>
>>>>> -----Inline Attachment Follows-----
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> postgis-users mailing list
>>>>> postgis-users at postgis.refractions.net<http://mc/compose?to=postgis-users@postgis.refractions.net>
>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> postgis-users mailing list
>>>>> postgis-users at postgis.refractions.net
>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>>>>
>>>>>
>>>>
>>>
>>> _______________________________________________
>>> postgis-users mailing list
>>> postgis-users at postgis.refractions.net
>>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>>
>>>
>>
>> _______________________________________________
>> postgis-users mailing list
>> postgis-users at postgis.refractions.net
>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>
>>
>
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20120219/ca28b719/attachment.html>