[GRASS-dev] [GRASS-user] r.to.vect stats

[Vaclav Petras]

On Fri, May 5, 2017 at 3:43 AM, Moritz Lennert <mlennert at club.worldonline.be
> wrote:

> Markus is probably right: working with an intermediate file solves all
> this and I don't think it would imply a serious speed decrease.
>

I have tested r.vect.stats (with pipe) and compared it with pipe in shell -
the time is the same (expected behavior). Here is the test with pipe in
shell for 19693050 (almost 20 million) points:

time sh -c "v.out.ascii input=points layer=-1 format=point sep=pipe |
r.in.xyz input=- output=bin2 method=mean sep=pipe --o"
real    1m5.982s
user    1m39.120s
sys    0m1.740s

And here is the same test but using temporary file (similar result with and
without SSD):

time sh -c "v.out.ascii input=points layer=-1 format=point sep=pipe
output=tmp.txt --o; r.in.xyz input=tmp.txt output=bin3 method=mean sep=pipe
--o"
real    1m32.933s
user    1m30.472s
sys    0m1.796s

According to this, using temporary file requires 1.4 more time.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-dev/attachments/20170505/c1b440d5/attachment.html>