[GRASS5] v.in.ascii

[Helena Mitasova]

I am trying to find a solution to the point data handling problem - we 
are currently stuck
at about 3mil pts on average PC.

I was able to read over 6mil points after commenting out Vect_build in 
v.in.ascii/in.c
I was able to display them but that was all I could do, all other 
commands
require level 2 (topology).
Several modules give error - cannot open on level 2 - (e.g. v.info, 
v.to.rast)
some give a more helpful message - you need to run v.build (but that 
can lead
to swapping and freezing of machine if the data set is over 3mil pts).

I am trying to get the following modules work with point data without 
topology:
v.info, v.to.rast, v.surf.rst - for none of these topology for points 
is needed (at least I guess).
Does anybody have any hints how to do that? I don't see an easy way how 
to separate
treatment of independent points from that of lines.
I have tried to find some documentation about what exactly is in the 
coor file and topo file -
is there a description available somewhere?

thanks,

helena



On Jun 30, 2005, at 5:53 AM, Radim Blazek wrote:

> On 6/30/05, Hamish <hamish_nospam at yahoo.com> wrote:
>>> There is no leak in the library, the memory is not released by 
>>> default
>>> because it takes long time. You can force the library to release the
>>> memory if you want using Vect_set_release_support().
>
> No, to build topology you need spatial index for all points.
>
> Radim
>
>> Perhaps then a clean & non-invasive solution is to add a check to run
>> Vect_set_release_support() once after each 100,000 points or so? or 
>> after
>> each time the allocation grows to > 50~100mb.
>>
>>
>> Can you safely run Vect_set_release_support() mid process?
>>
>>
>> ?
>> Hamish
>>
>
>
Helena Mitasova
Dept. of Marine, Earth and Atm. Sciences
1125 Jordan Hall, NCSU Box 8208,
Raleigh NC 27695
http://skagit.meas.ncsu.edu/~helena/