[Qgis-user] Best practice, database vs WFS

Richard Duivenvoorde rdmailings at duif.net
Thu Jan 7 02:01:17 PST 2021

On 1/7/21 10:40 AM, Paul Wittle wrote:
> It looks like there are a number of possible solutions but at present it seems we will have to consider the options again and change our strategy accordingly. I have already created a python plugin for QGIS which reads metadata from GeoServer and populates a list of available layers. I had been developing the concept of then allowing the user to choose which format to use (direct access to db, WFS or WMS). At present this appears to be the best approach but we may need to redesign our ideas for the database setup accordingly if users will be preferring direct db layers in QGIS.

Or do both?

First create a proper Postgis database(model), which QGIS users can create/edit etc directly from intern.

Publish (views) of this database with Geoserver to display outside of the office
Make tables in this database editable and use Geoserver WFS-T to edit in the field (even outside of the office), using leaflet/openlayers apps.
WFS can also do Geojson.

And: *who* is editing (aka fiddling with your data), what? Like: are some specialists editing cadastral parcels (say in QGIS), or is it the public digitizing (roughly) carbage in a city.
I know of people doing this with a WFS with, say 100000 points with animal location, and doing fine.
It also depends on the data: millions of records, or (10)thousands.
Points are easier to handle then complex polygons (or even worse: curves...) in WFS-T

And to make it even more complex: there are also small QGIS-based apps like 'Input' and 'QField', which (if I am correct?) can be used offline and then sync there data later.

Keep it modular.

Choices, choices, choices. Not my strongest thing...


Richard Duivenvoorde

More information about the Qgis-user mailing list