[postgis-users] problem with raster2pgsql

Bborie Park bkpark at ucdavis.edu
Thu Mar 8 11:32:17 PST 2012


Generally, in Linux you would only have one instance of the GDAL library 
installed.  This one instance would be what any dependent programs would 
call.

Granted, I always build my own GIS stack (PROJ.4, GEOS, PostgreSQL, 
GDAL, PostGIS, QGIS) so I am able to have absolute control over how 
things are built.

In Windows and possibly OSX, that may not be the case.  QGIS may provide 
it's own copy of the GDAL library while PostGIS is providing its own one.

As for GDAL 1.9, everything is the same.

-bborie

On 03/08/2012 10:41 AM, Pierre Racine wrote:
> So I understand that, on Linux, the GDAL library used by PostGIS is the same as the one used by the GDAL utilities and QGIS. Which is not the case on Windows where PostGIS has its own GDAL dll. Right?
>
> What if you test with GDAL 1.9?
>
> Pierre
>
>> -----Original Message-----
>> From: postgis-users-bounces at postgis.refractions.net [mailto:postgis-users-
>> bounces at postgis.refractions.net] On Behalf Of Bborie Park
>> Sent: Thursday, March 08, 2012 1:38 PM
>> To: PostGIS Users Discussion
>> Subject: Re: [postgis-users] problem with raster2pgsql
>>
>> I just double-checked the images on QGIS 1.7.3 using the latest GDAL trunk
>> (r24093) with the PostGIS Raster plugin.
>>
>> I've attached screenshots in QGIS.
>>
>> On Thu, Mar 8, 2012 at 10:29 AM, Bborie Park<bkpark at ucdavis.edu>  wrote:
>>> Cheryl,
>>>
>>> I took a look at your adjusted images and they fine by me.  I loaded
>>> the PNG files into PostGIS Raster and ran an ST_Histogram on the
>>> aspect image to see if the data is there.
>>>
>>> I called raster2pgsql with the following...
>>>
>>> raster2pgsql -t 50x50 -I -C -Y Aspect.png aspect>  aspect.sql
>>>
>>> My bet is that there is something wrong with QGIS, which uses GDAL to
>>> access PostGIS raster.  There is another email thread related to
>>> PostGIS raster access using GDAL and problems arising from that.
>>>
>>> -bborie
>>>
>>> On Thu, Mar 8, 2012 at 3:07 AM, cheryl buckley<cheryl.bckl at gmail.com>
>> wrote:
>>>> Hi bborie and Pierre,
>>>>
>>>> I am not too sure if what I did was a hundred percent correct. I was
>>>> thinking perhaps I am suppose to make a setting in QGIS to obtain a
>>>> corrected image, and not alter the "-t 100x100" in the raster2pgsql
>>>> command line, i'm not too sure. But, non-the-less find attached zip
>>>> folders (screenshots) of my images I generated. The first set is the
>>>> faulty images with the "-t 100x100" and the second set is the images
>>>> generated with a "-t 300x300".  Please note that the hillshade image
>>>> in this case also came out faulty. I also made screenshots of the
>>>> histograms of the images. Perhaps this will now shed some light.
>>>> Please do confirm if you were able to unzip the images and view them.
>>>>
>>>> Regards
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Thu, Mar 8, 2012 at 10:14 AM, Pierre Racine
>>>> <Pierre.Racine at sbf.ulaval.ca>  wrote:
>>>>>> Yes.  All tiles of a raster will have the same dimensions.  So an
>>>>>> edge tile would have the same dimensions as every other tile, just
>>>>>> that some of the tiles have no data.
>>>>>
>>>>> Probably no link with the bug being discussed but I'm curious. What happen
>> when the raster do not have a nodata value defined and you have to pad some
>> tiles?
>>>>>
>>>>> Pierre
>>>>> _______________________________________________
>>>>> postgis-users mailing list
>>>>> postgis-users at postgis.refractions.net
>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>>>
>>>>
>>>>
>>>> --
>>>> Greetings
>>>> Cheryl
>>>>
>>>> _______________________________________________
>>>> postgis-users mailing list
>>>> postgis-users at postgis.refractions.net
>>>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>>>
>>>
>>>
>>>
>>> --
>>> Bborie Park
>>> Programmer
>>> Center for Vectorborne Diseases
>>> UC Davis
>>> 530-752-8380
>>> bkpark at ucdavis.edu
>>
>>
>>
>> --
>> Bborie Park
>> Programmer
>> Center for Vectorborne Diseases
>> UC Davis
>> 530-752-8380
>> bkpark at ucdavis.edu
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>

-- 
Bborie Park
Programmer
Center for Vectorborne Diseases
UC Davis
530-752-8380
bkpark at ucdavis.edu



More information about the postgis-users mailing list