[Qgis-developer] Processing - current status of "in-between" layers while running models

[kimaidou]

Hi all !

I was wondering what was the current status for layers created as ouput of
algs and used as input in child algs ? I remember reading some new feature
in this area, but I cannot remember what was the conclusion. Is there some
options allowing to do the following :

* use memory layers for all intermediate layers -> could be very fast, but
there should be some way to prevent bloating the computer if the input data
is too big ( the user would prefer to avoid memory layer if it leads to
swap )

* use a unique spatialite database to store the intermediate layers. In
this case
  - create automatically spatial indexes ( it is fast )
  - use rasterlite to store raster layers ? Or use a kind of foreign data
wrapper for sqlite ?

* For both options, obviously keep the long field names along the process,
and the complex data types ( and do not restrict to the 4 datatypes
available in DBF )

* I am aware that the use of Grass or other rich algs provider (Saga,
Taudem) can lead to using their internal format. What are the possibilities
and limits in this case ? For example for vector, Grass needs to import the
data, but what about for raster like GeoTiffs ?

Thanks in advance for your sharing your opinions and knowledge on this.

Regards,
Michaël
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20160516/e61d221f/attachment-0001.html>