[Qgis-user] Support for Mesh layers (netCDF) on Postgis/Postgresql: is it possible?

[Jorge Gustavo Rocha]

Hi Saber,

Thank you for your feedback. You are an expert on the subject and I 
really appreciate your comments.

Postgresql is probably not the best option in terms of performance. But 
it is a practical solution to centralize all information in a single place.

I'm using postgis raster support, for example, which is not faster than 
files, but it is ok for my use cases. I have all the spatial data in one 
place, where I can manage authentication and authorizations quite well. 
The end user will use a cached WMTS/TMS service (MapProxy on top of QGIS 
Server), and the performance will be ok. I want to use QGIS Server to 
provide data in netCDF files as temporal layers. The end user will use 
cached data and will not notice any performance issues. The cache can be 
created once, since the data is always in the past. That's my use case: 
I'm not processing the data or making any processing. Just providing 
maps/tiles.

GDAL and QGIS works very well with Postgresql and users are familiar 
with Postgresql (setup and usage).

How difficult would be to have netCDF data in Postgresql?

1) We should have a way to store the data (and metadata) on the database 
(like we do with rasters, with raster2pgsql).

2) We need to improve QgsMeshLayer to read from a PG uri, like:

uri_config = {
     'service':'atlantico',
     'schema':'public',
     'table':'az1_hs_2023101200_03',
     'geometrycolumn':'rast'
}
uri_config = {key:val for key, val in uri_config.items() if val is not None}
md = QgsProviderRegistry.instance().providerMetadata('mdal')
uri = QgsDataSourceUri(md.encodeUri(uri_config))
mesh_layer = QgsMeshLayer(uri.uri(False), "layer_name", "mdal")

If you think this does not make sense, you are probably right :-) I'm 
quite new to netCDF data and I just want to use the tools I'm used to 
have with this format.

Regards,

Jorge

On 14/12/23 10:13, Saber Razmjooei wrote:
> Hi Jorge,
>
> What is your motivation for storing mesh data in PG? If you want to 
> have fast access to large central datasets, PG is probably not a good 
> storage option. Best to investigate Zarr (https://zarr.dev/) which is 
> a cloud optimised format for large multidimensional datasets. We need 
> to first implement it in MDAL for QGIS to be able to load the data 
> effectively.
>
> Another option to serve large NetCDF/GRIB data is to set up an EDR 
> (https://ogcapi.ogc.org/edr/overview.html) service which we are adding 
> support for it in QGIS.
>
> Kind regards
> Saber
>
>
> On Wed, 13 Dec 2023 at 19:36, Jorge Gustavo Rocha via QGIS-User 
> <qgis-user at lists.osgeo.org> wrote:
>
>     Hi,
>
>     I'm playing with netCDF files. I add them to QGIS using the mdal
>     provider (mesh layer). Reading netCDF data from files works very well.
>
>     My question is: is it possible to read netCDF data from
>     Postgis/Postgresql?
>
>     I can store netCDF files in Postgis using raster2pgsql, but the
>     data is
>     stored as raster and I can only open it as raster in QGIS and not as
>     mesh. Is works, but not as good as the mesh support.
>
>     I'm storing it, using something like this:
>
>     raster2pgsql -s 4326 -d -I -C -M AZ1_HS_2023101200.nc -F
>     public.az1_hs_2023101200 | psql service=atlantico
>
>     Thanks in advance,
>
>     Jorge Gustavo
>
>     _______________________________________________
>     QGIS-User mailing list
>     QGIS-User at lists.osgeo.org
>     List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
>     Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
>
>
>
> -- 
> Saber Razmjooei
> www.lutraconsulting.co.uk <http://www.lutraconsulting.co.uk>
> +44 (0)7568 129733
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20231214/f73928b6/attachment.htm>