[GRASS-dev] Import PostGIS vector layer into SQLite database

[Luí­s Moreira de Sousa]

Hi all,

I am working with vector layers (points) composed of more than 100 000 geometries. These layers are originally in a PostGis database and I import them with v.in.ogr, using a formulation like:

v

.

in

.

ogr

\

input

=

"PG:host=127.0.0.1 dbname=gisdb user=user password=secret"

\

layer

=

schema

.

points

\

output

=

points

\

-

o

--overwrite

This command is successful, however, operations with the resulting points layer are remarkably slow. For instance, this layer must be later loaded in R and that operation alone is taking over 3 hours.

After some back-and-forth testing, I identified the problem: v.in.ogr is not saving its output in the internal GRASS SQLite database. Knowing this, I can create a copy of the original output, that is automatically inserted into the SQLite database, e.g.:

g.copy vector=points,points_mirror

Importing the points_mirror layer to R takes 3 minutes instead of 3 hours, for instance.

While mirroring vectors imported with v.in.ogr solves the issue, it is an inelegant solution that also takes up unnecessary space. I would prefer to instruct v.in.ogr somehow to force it to use the SQLite database. I could not find any reference in the manual, is it possible to do?

Thank you.

--
Luís Moreira de Sousa
Email: luis.de.sousa at protonmail.ch
RingID: ring:7ca91d83f4f9dec82fec9f1144b8e5c1ef2a110c
URL: https://ldesousa.github.io

Sent with [ProtonMail](https://protonmail.com) Secure Email.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-dev/attachments/20190327/aba83093/attachment-0001.html>