[GRASS-user] importing missing values via v.in.ascii for subsequent interpolation

GRASS GRASS user
1

Hello,

I am trying to import ASCII files with v.in.ascii for subsequent interpolation via v.surf.idw. The problem is that if I put NULL for missing values, v.in.ascii complains about the column being of type DOUBLE and fails. On the other hand if I put missing values as just '', as suggested at
http://www.intevation.de/rt/webrt?serial_num=2763&display=History
then v.surf.rst seems to use those missing values as ZERO values in computing the weighted average.

I tried with something like

  v.db.update map=myMap column=myColumn value="NULL" \
    where=myColumn=''

But it did not work.

Does anybody have a way around this?

Regards,

Luigi

2

On Tuesday 14 November 2006 18:40, Luigi Ponti wrote:

Hello,

I am trying to import ASCII files with v.in.ascii for subsequent
interpolation via v.surf.idw. The problem is that if I put NULL for
missing values, v.in.ascii complains about the column being of type
DOUBLE and fails. On the other hand if I put missing values as just '',
as suggested at
http://www.intevation.de/rt/webrt?serial_num=2763&display=History
then v.surf.rst seems to use those missing values as ZERO values in
computing the weighted average.

I tried with something like

  v.db.update map=myMap column=myColumn value="NULL" \
    where=myColumn=''

But it did not work.

Does anybody have a way around this?

Regards,

Luigi

hi.

how about this:

1. change any nodata values in your input file to an empty string
2. v.in.ascii in=infile sep=sep ... out=outvect
3. double check that the nodata values were imported correctly, i.e missing
from the attribute table
4. v.extract in=outvect where="column = ..." out=vect_no_null
5. v.surf.idw vect_no_null

note that depending on the DB backend you are using, you may need to tweak the
above. another approach might be to code your nodata values as -9999999 or
something like that, and then use that to select only features with 'real'
data.

cheers,

--
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341

3

Dylan Beaudette wrote:

On Tuesday 14 November 2006 18:40, Luigi Ponti wrote:
  

Hello,

I am trying to import ASCII files with v.in.ascii for subsequent
interpolation via v.surf.idw. The problem is that if I put NULL for
missing values, v.in.ascii complains about the column being of type
DOUBLE and fails. On the other hand if I put missing values as just '',
as suggested at
http://www.intevation.de/rt/webrt?serial_num=2763&display=History
then v.surf.rst seems to use those missing values as ZERO values in
computing the weighted average.

I tried with something like

  v.db.update map=myMap column=myColumn value="NULL" \
    where=myColumn=''

But it did not work.

Does anybody have a way around this?

Regards,

Luigi

hi.

how about this:

1. change any nodata values in your input file to an empty string 2. v.in.ascii in=infile sep=sep ... out=outvect
3. double check that the nodata values were imported correctly, i.e missing from the attribute table
4. v.extract in=outvect where="column = ..." out=vect_no_null
5. v.surf.idw vect_no_null

note that depending on the DB backend you are using, you may need to tweak the above. another approach might be to code your nodata values as -9999999 or something like that, and then use that to select only features with 'real' data.

cheers,

Thanks Dylan!
The -9999999 approach works like beauty.
Luigi