[GRASS-user] g.region -a: what's it supposed to do?

GRASS GRASS user
1

Hello,

I just got caught by a surprising g.region behavior. Can someone clarify what the "-a" flag is supposed to do? From the docs, "Align region to resolution (default = align to bounds, works only for 2D resolution)".

I have a defined region on a 1000 m grid. The two are defined together and do "align" as seen by the first set of commands (here "align" means the bounds and the resolution match in both ns and ew directions, and there are no fractional values). In this scenario, why is the application of the "-a" flag shifting the boundaries? What is being aligned to what?

My previous understanding is that if I set e,w,n,s AND rows,cols AND res, and they don't all agree, then res takes precedence if -a is set, and otherwise res is subject to the other six paramaters. Alternatively, if just setting e,w,n,s and res, and the e,w res does not match the n,s res, then setting -a would adjust things. But when everything agrees, why is -a shifting the bounds?

grass -c EPSG:3413 G_tmp

g.region w=-638956 e=856044 s=-3354596 n=-655596 res=1000 -p
eval $(g.region -g)
echo $(( ($n - $s)/1000 ))
echo $(( ($e - $w)/1000 ))

# why does this shift things?
g.region w=-638956 e=856044 s=-3354596 n=-655596 res=1000 -pa

I don't see previous questions about this when searching "g.region resolution align bounds site:lists.osgeo.org/pipermail/grass-user/" on Google.

Thanks,

   -k.

2

Hello again,

Please disregard the previous message. I've read the docs again and "even multiples of the resolution" is the key part. I get it now, and why I'm seeing the behavior from the commands below.

  -k.

On 2020-11-27 at 04:13 -08, Ken Mankoff <mankoff@gmail.com> wrote...

My previous understanding is that if I set e,w,n,s AND rows,cols AND
res, and they don't all agree, then res takes precedence if -a is set,
and otherwise res is subject to the other six paramaters.
Alternatively, if just setting e,w,n,s and res, and the e,w res does
not match the n,s res, then setting -a would adjust things. But when
everything agrees, why is -a shifting the bounds?

grass -c EPSG:3413 G_tmp

g.region w=-638956 e=856044 s=-3354596 n=-655596 res=1000 -p
eval $(g.region -g)
echo $(( ($n - $s)/1000 ))
echo $(( ($e - $w)/1000 ))

# why does this shift things?
g.region w=-638956 e=856044 s=-3354596 n=-655596 res=1000 -pa

I don't see previous questions about this when searching "g.region resolution align bounds site:lists.osgeo.org/pipermail/grass-user/" on Google.

Thanks,

   -k.

3

Hi Ken,

On Sat, Nov 28, 2020 at 7:55 PM Ken Mankoff <mankoff@gmail.com> wrote:

Hello again,

Please disregard the previous message. I've read the docs again and "even multiples of the resolution" is the key part. I get it now, and why I'm seeing the behavior from the commands below.

If you have suggestions how to improve the documentation: please don't
hesitate to edit:

https://github.com/OSGeo/grass/blob/master/general/g.region/g.region.html
(edit button, upper right side; requires a GitHub login).

When saving, the system will suggest it as a pull request which can
then be reviewed by anyone and eventually merged (you will know all
this but I just write it to invite all community members to easily
contribute).

cheers,
Markus