Hey Standa,

It looks like the high to low output range was causing issues with the code. So, please try r8393 where I've added a check for the range direction.

My test query of

{{{
WITH foo AS (
        SELECT ST_AddBand(
                ST_MakeEmptyRaster(10, 10, 0., 0., 1., -1., 0, 0, 0),
                '8BUI', 100., 0
        ) AS rast
)
SELECT
        ST_Reclass(rast, 1, '1-100]:100-1', '8BUI', 0)
FROM foo
}}}

does as expected where the value 100 is set to 1 in the output raster.

-bborie

On 12/13/2011 12:11 AM, Stanislav Vanecek wrote:
Hi Bborie
The SQL is
SELECT 
ST_AsTIFF(ST_AddBand(ST_AddBand(ST_AddBand(ST_MakeEmptyRaster(rast),ST_Reclass(rast,1,
 '0-2115):255-0'::text, '8BUI'::text,0)),ST_Reclass(rast,1, 
'0-2115):255-0'::text, '8BUI'::text,0)),ST_Reclass(rast,1, 
'0-2115):255-0'::text, '8BUI'::text,0))
And in this case the result is white rectangle
What I am doing wrong?
Best regards
Standa


-----Original Message-----
From: postgis-users-boun...@postgis.refractions.net 
[mailto:postgis-users-boun...@postgis.refractions.net] On Behalf Of Bborie Park
Sent: Monday, December 12, 2011 4:41 PM
To: PostGIS Users Discussion
Subject: Re: [postgis-users] ST_Reclass - oposit direction in reclass

On Mon, Dec 12, 2011 at 6:42 AM, Stanislav Vanecek<s.vane...@dhi.cz>  wrote:
Hi
  I am using the ST_Reclass to create 3-band TIFF (for graphical
presentation) from 1 band grid 32BF (using algorithm described in 32BF
http://postgis.refractions.net/documentation/manual-svn/RT_ST_Reclass.ht
ml)
The problem is, then for the mapping both intervals in (reclassexpr
text) must be increasing  - values from small to high is mapped to
colors from "small  to high" = (black to white for example) - and I like
to have them in opposite order.
I am working with the big rasters (2000 x 2000 cells). In this case the
ST_Reclass is already not very fast - and when I try to use map algebra
it was very, very slow.
Did someone have idea how to speed up?

Standa,

Can you post your SQL?  I don't think there is anything restricting
you from having your intervals from high to low instead of low to
high, but I could be wrong.  ST_Reclass should be faster than
ST_MapAlgebra as all evaluations are done in C instead of going back
and forth with the PostgreSQL server.

-bborie


--
Bborie Park
Programmer
Center for Vectorborne Diseases
UC Davis
530-752-8380
bkp...@ucdavis.edu
_______________________________________________
postgis-users mailing list
postgis-users@postgis.refractions.net
http://postgis.refractions.net/mailman/listinfo/postgis-users

Reply via email to