[GRASS-user] i.cluster fails and how to create a working attribute data table

GRASS GRASS user
1

Hi Folks,

Once again, I got some problems with the windows version of GRASS.

When I want to do an unsupervised classification, the output of i.cluster
reads:

"G_calloc: unable to allocate 1000000001 * 8 bytes at alloc_cell.c:128"

The command I entered reads as follows:

i.cluster group=Landsat7@PERMANENT subgroup=LS7_SUB sigfile=LS7_SIG
classes=20 iterations=30 convergence=98.0 separation=0.0 min_size=17
reportfile=Rep

The other problem concerning the creation of vector features is, that i
can't enter any data, as the error output reads:

Cannot open database
'E:/Damokles_BACKUP/Backup/Wichtig/Geomatik/GIS2/Präsentation/GIS
DataBase/Deutschland/Ruhrgebiet/dbf/' by driver 'dbf'

So how do I create prober attribute data tables linked to a certain vector
layer?

Kind regards,
Dominik
--
View this message in context: http://n2.nabble.com/i.cluster-fails-and-how-to-create-a-working-attribute-data-table-tp3301468p3301468.html
Sent from the Grass - Users mailing list archive at Nabble.com.

2

Dominik Wiedner wrote:

Hi Folks,

Once again, I got some problems with the windows version of GRASS.

When I want to do an unsupervised classification, the output of i.cluster
reads:

"G_calloc: unable to allocate 1000000001 * 8 bytes at alloc_cell.c:128"
  

This is a request for more than 7GB of memory, unlikely that this is needed for a Landsat scene. Do the current region settings match the raster to be processed?

The command I entered reads as follows:

i.cluster group=Landsat7@PERMANENT subgroup=LS7_SUB sigfile=LS7_SIG
classes=20 iterations=30 convergence=98.0 separation=0.0 min_size=17
reportfile=Rep

The other problem concerning the creation of vector features is, that i
can't enter any data, as the error output reads:

Cannot open database
'E:/Damokles_BACKUP/Backup/Wichtig/Geomatik/GIS2/Präsentation/GIS
DataBase/Deutschland/Ruhrgebiet/dbf/' by driver 'dbf'

A wild guess: it may be the "ä" giving trouble.

Markus M

3

Markus Metz-2 wrote:

This is a request for more than 7GB of memory, unlikely that this is
needed for a Landsat scene. Do the current region settings match the
raster to be processed?

Hi Markus,

This problem is solved, but another appears!

When I want to execute i.maxlik, the output reads, that the file with the
signatures can't be read.

Kind regards,
Dominik

--
View this message in context: http://n2.nabble.com/i.cluster-fails-and-how-to-create-a-working-attribute-data-table-tp3301468p3301640.html
Sent from the Grass - Users mailing list archive at Nabble.com.

4

Dominik Wiedner wrote:

When I want to execute i.maxlik, the output reads, that the file with the
signatures can't be read.
  

Now I'm lost and can only ask stupid questions: Can you read the signature file with a standard text editor and do the contents make sense? No offence, but did you specify the right file name (no spelling mistakes)?

Any further answers are probably more helpful if you could post the commands you used.

Markus M