A couple of months ago I sent a few messages regarding a memorry error with r.to.vect (it was allocating to many information and data). My question is, sicne r.to.vect was not updated, is there any raster or vector lib update that had any impact on this?
Thanks
Antonio Rocha
__________ Information from ESET NOD32 Antivirus, version of virus signature database 5977 (20110323) __________
The message was checked by ESET NOD32 Antivirus.
2011/3/23 António Rocha <antonio.rocha@deimos.com.pt>:
A couple of months ago I sent a few messages regarding a memorry error with
r.to.vect (it was allocating to many information and data). My question is,
sicne r.to.vect was not updated, is there any raster or vector lib update
that had any impact on this?
Did you post a test case? Best done with the North Carolina sample
data set. Otherwise it is hard to replicate...
Markus
Hello Markus
Unfortunely, I get this error when I run other datasets. In this case, a landCover with all CELLS and spatial resolution of 30. The problem is that North carolina sample is too small.
r.to.vect -v input=FULL@PERMANENT output=FULL feature=area
I get this:
Default driver / database set to:
driver: dbf
database: $GISDBASE/$LOCATION_NAME/$MAPSET/dbf/
Extracting areas...
Building topology for vector map <PT_National_2000_FULL>...
R4243000ERROR: G_realloc: unable to allocate 16976004 bytes at struct_alloc.c:133
It has created a vector but I cannot display and I cannot use v.info (it says that I need to build topology).
And If I try to build topology I get and error and before this warning:
Coor files of vector map <PT_National_2000_FULL@PERMANENT> is larger than it should be (241464079 bytes excess)
So my questions are:
1- Is any fix for this?
2- If not, is there any defined limit for a raster in order to be converted to vector?
Thanks
Antonio
Markus Neteler wrote:
2011/3/23 António Rocha <antonio.rocha@deimos.com.pt>:
A couple of months ago I sent a few messages regarding a memorry error with
r.to.vect (it was allocating to many information and data). My question is,
sicne r.to.vect was not updated, is there any raster or vector lib update
that had any impact on this?
Did you post a test case? Best done with the North Carolina sample
data set. Otherwise it is hard to replicate...Markus
__________ Information from ESET NOD32 Antivirus, version of virus signature database 5995 (20110329) __________
The message was checked by ESET NOD32 Antivirus.
2011/3/29 António Rocha <antonio.rocha@deimos.com.pt>:
Hello Markus
Unfortunely, I get this error when I run other datasets. In this case, a
landCover with all CELLS and spatial resolution of 30. The problem is that
North carolina sample is too small.
r.to.vect -v input=FULL@PERMANENT output=FULL feature=area
I get this:
Default driver / database set to:
driver: dbf
database: $GISDBASE/$LOCATION_NAME/$MAPSET/dbf/
Extracting areas...
Building topology for vector map <PT_National_2000_FULL>...
R4243000ERROR: G_realloc: unable to allocate 16976004 bytes at
struct_alloc.c:133
You mean that it is too big 
The file in question is lib/vector/diglib/struct_alloc.c
Please try GRASS 7 instead of GRASS 6 which offers large file support
for vector data (to some extent).
You need to configure it with this options
--enable-largefile enable support for large files (LFS)
and compile it. Or use the weekly Linux binary snapshot which
has LFS enabled.
Markus
Hello Markus
Thanks.
Just one final question: Is there a way to know which is the limit for vector? My idea is to check if raster, to be converted, is too big to be converted before I try to convert?
THanks
Markus Neteler wrote:
Hi António
(cc Martin)you can get winGRASS 7 here:
http://wingrass.fsv.cvut.cz/grass70/but I dunno if LFS is enabled in that built. Maybe Martin
can tell us.cheers
Markus2011/3/29 António Rocha <antonio.rocha@deimos.com.pt>:
Hi Markus
The problem is that I'm running in Windows and in GRASS6.4.0stable release.
So, there is no solution for Windows?
Thanks
AntonioMarkus Neteler wrote:
2011/3/29 António Rocha <antonio.rocha@deimos.com.pt>:
Hello Markus
Unfortunely, I get this error when I run other datasets. In this case, a
landCover with all CELLS and spatial resolution of 30. The problem is
that
North carolina sample is too small.
r.to.vect -v input=FULL@PERMANENT output=FULL feature=area
I get this:
Default driver / database set to:
driver: dbf
database: $GISDBASE/$LOCATION_NAME/$MAPSET/dbf/
Extracting areas...
Building topology for vector map <FULL>...
R4243000ERROR: G_realloc: unable to allocate 16976004 bytes at
struct_alloc.c:133You mean that it is too big
The file in question is lib/vector/diglib/struct_alloc.cPlease try GRASS 7 instead of GRASS 6 which offers large file support
for vector data (to some extent).You need to configure it with this options
--enable-largefile enable support for large files (LFS)and compile it. Or use the weekly Linux binary snapshot which
has LFS enabled.Markus
__________ Information from ESET NOD32 Antivirus, version of virus signature database 5995 (20110329) __________
The message was checked by ESET NOD32 Antivirus.
2011/3/29 António Rocha <antonio.rocha@deimos.com.pt>:
Hello Markus
Unfortunely, I get this error when I run other datasets. In this case, a
landCover with all CELLS and spatial resolution of 30. The problem is that
North carolina sample is too small.
r.to.vect -v input=FULL@PERMANENT output=FULL feature=area
I get this:
Default driver / database set to:
driver: dbf
database: $GISDBASE/$LOCATION_NAME/$MAPSET/dbf/
Extracting areas...
Building topology for vector map <PT_National_2000_FULL>...
R4243000ERROR: G_realloc: unable to allocate 16976004 bytes at
struct_alloc.c:133It has created a vector but I cannot display and I cannot use v.info (it
says that I need to build topology).
And If I try to build topology I get and error and before this warning:
Coor files of vector map <PT_National_2000_FULL@PERMANENT> is larger than it
should be (241464079 bytes excess)So my questions are:
1- Is any fix for this?
2- If not, is there any defined limit for a raster in order to be converted
to vector?
Since this is an out-of-memory error, you could try grass 7 and in your terminal
export GRASS_VECTOR_LOW_MEM=1
before running r.to.vect. This will reduce memory requirements while
building topology. Large File Support is probably not needed here, but
nowadays a good thing to have, just in case.
Markus M
Markus Neteler wrote:
2011/3/23 António Rocha <antonio.rocha@deimos.com.pt>:
A couple of months ago I sent a few messages regarding a memorry error
with
r.to.vect (it was allocating to many information and data). My question
is,
sicne r.to.vect was not updated, is there any raster or vector lib update
that had any impact on this?Did you post a test case? Best done with the North Carolina sample
data set. Otherwise it is hard to replicate...Markus
Hi Markus
Thanks for the feedback but I am running in WinGRASS. Just two questions:
- Is it possible to have Long File Support in wiNGRASS 6.4?
- Since I'm running this in a Script, how can I define " export GRASS_VECTOR_LOW_MEM=1 " in a Python Script?
Thanks
Antonio
Markus Metz wrote:
2011/3/29 António Rocha <antonio.rocha@deimos.com.pt>:
Hello Markus
Unfortunely, I get this error when I run other datasets. In this case, a
landCover with all CELLS and spatial resolution of 30. The problem is that
North carolina sample is too small.
r.to.vect -v input=FULL@PERMANENT output=FULL feature=area
I get this:
Default driver / database set to:
driver: dbf
database: $GISDBASE/$LOCATION_NAME/$MAPSET/dbf/
Extracting areas...
Building topology for vector map <PT_National_2000_FULL>...
R4243000ERROR: G_realloc: unable to allocate 16976004 bytes at
struct_alloc.c:133It has created a vector but I cannot display and I cannot use v.info (it
says that I need to build topology).
And If I try to build topology I get and error and before this warning:
Coor files of vector map <PT_National_2000_FULL@PERMANENT> is larger than it
should be (241464079 bytes excess)So my questions are:
1- Is any fix for this?
2- If not, is there any defined limit for a raster in order to be converted
to vector?Since this is an out-of-memory error, you could try grass 7 and in your terminal
export GRASS_VECTOR_LOW_MEM=1
before running r.to.vect. This will reduce memory requirements while
building topology. Large File Support is probably not needed here, but
nowadays a good thing to have, just in case.Markus M
Markus Neteler wrote:
2011/3/23 António Rocha <antonio.rocha@deimos.com.pt>:
A couple of months ago I sent a few messages regarding a memorry error
with
r.to.vect (it was allocating to many information and data). My question
is,
sicne r.to.vect was not updated, is there any raster or vector lib update
that had any impact on this?Did you post a test case? Best done with the North Carolina sample
data set. Otherwise it is hard to replicate...Markus
__________ Information from ESET NOD32 Antivirus, version of virus signature database 6000 (20110330) __________
The message was checked by ESET NOD32 Antivirus.
Hi,
2011/3/30 António Rocha <antonio.rocha@deimos.com.pt>:
GRASS_VECTOR_LOW_MEM=1 " in a Python Script?
os.environ['GRASS_VECTOR_LOW_MEM'] = 1
Note that GRASS 7 is required [1]
Martin
[1] http://wingrass.fsv.cvut.cz/grass70/
--
Martin Landa <landa.martin gmail.com> * http://geo.fsv.cvut.cz/~landa