[gdal-dev] reading postgis raster in mode 2 error

[Jorge Arévalo]

On Mon, Aug 29, 2011 at 12:15 PM, Ricardo Filipe Soares Garcia da
<ricardo.garcia.silva at gmail.com> wrote:
> 2011/8/26 Jorge Arévalo <jorge.arevalo at deimos-space.com>:
>> ---------- Forwarded message ----------
>> From: Ricardo Filipe Soares Garcia da <ricardo.garcia.silva at gmail.com>
>> Date: Wed, Aug 24, 2011 at 6:07 PM
>> Subject: reading postgis raster in mode 2 error
>> To: gdal-dev at lists.osgeo.org
>>
>>
>> Hi list
>>
>> I am trying out the postgis raster driver.
>>
>> I'm running Ubuntu 11.04 with gdal 1.8 from the ubuntugis repository.
>> As for postgis I've compiled a snapshot of postgis 2.0 from svn, as
>> offered on the postgis website.
>>
>> Following the directions in the driver info page[1], I've been trying
>> to get a gdalinfo on the katrina.tif file. When I run it using
>> mode='1' everything goes as expected and I get info on each of the
>> tiles.
>>
>> but when using mode='2' I am getting this error:
>>
>> gdalinfo -mm -stats -checksum "PG:host='localhost' dbname='gis_testes'
>> user='gisuser' password='resusig' table='katrina' mode='2'"
>>
>>
>> ERROR 1: Error, the                     ONE_RASTER_PER_TABLE mode
>> can't be applied if the raster                     rows don't have the
>> same metadata for band 1
>> gdalinfo failed - unable to open 'PG:host='localhost'
>> dbname='gis_testes' user='test_user' password='test_pass'
>> table='katrina' mode='2''.
>>
>>
>> When importing the file into postgis I ran the following command:
>>
>> /usr/lib/postgresql/8.4/bin/raster2pgsql.py -r ~/Downloads/katrina.tif
>> -t katrina -k 64x64 -o katrina.sql -s 4326 -I
>>
>>
>> Thanks for your help
>>
>> [1] - http://trac.osgeo.org/gdal/wiki/frmts_wtkraster.html
>>
>> --
>> ___________________________ ___ __
>> Ricardo Garcia Silva
>>
>>
>> Hi Ricardo,
>>
>> That error is caused because this query
>>
>> select (foo.md).* from (select distinct st_bandmetadata(rast, n) as md
>> from katrina) as foo
>>
>> (being 'n' the band number).
>>
>> returns more than one result, and it shouldn't. Could you please
>> execute that query in a client and send me back the results?
>>
>> Best regards, and sorry for the inconvenience
>>
>> --
>> Jorge Arévalo
>> Internet & Mobility Division, DEIMOS
>> jorge.arevalo at deimos-space.com
>> http://es.linkedin.com/in/jorgearevalo80
>> http://mobility.grupodeimos.com/
>> http://gis4free.wordpress.com
>> http://geohash.org/ezjqgrgzz0g
>>
>
> Hi Jorge, list
>
> Running the query you suggest I get this output:
>
> gis_testes=> select (foo.md).* from (select distinct
> st_bandmetadata(rast, 1) as md from katrina) as foo;
>  pixeltype | hasnodatavalue | nodatavalue  | isoutdb | path
> -----------+----------------+--------------+---------+------
>  8BUI      | f              | -1.82132e-05 | f       |
>  8BUI      | f              | -1.81834e-05 | f       |
>  8BUI      | f              | -1.80478e-05 | f       |
>  8BUI      | f              | -1.80471e-05 | f       |
> (4 rows)
>
> Can it be that the nodatavalue is messing things up, since it is not
> the same for all the blocks?
>
> Thanks for your help
>
>
> --
> ___________________________ ___ __
> Ricardo Garcia Silva
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/gdal-dev
>

Hi,

Yes, it's a problem with metadata precission. I found this error in a
different part of the code. I'm not sure if is caused because changes
in PostGIS Raster code, but I'm aware of it. I applied a temporal
patch in r23006. Could you please update your GDAL copy?

Best regards,

-- 
Jorge Arévalo
Internet & Mobility Division, DEIMOS
jorge.arevalo at deimos-space.com
http://es.linkedin.com/in/jorgearevalo80
http://mobility.grupodeimos.com/
http://gis4free.wordpress.com
http://geohash.org/ezjqgrgzz0g