[Qgis-user] Using custom projections/transformations from a PROJ.4 init-file

[Andre Joost]

Am 12.01.2018 um 11:24 schrieb Kristian Evers:
> Hi,
>
> I am trying to define my own "custom projection" by using a PROJ.4
> init-file. The reason I want to do that is that I only want to
> maintain the projections/transformations that I need  in one place
> instead of many. I find the process a bit frustrating since at first
> glance it seems to be working, but then when I try to save it and
> close the custom projection dialog I am told that the definition is
> invalid. As it can be seen on the screenshot below, when I enter the
> PROJ.4 parameters - in this case
> "+init=C:\dev\NordicTransformations\resources\DK:UTM32N" - the test
> works fine but when I try to save it, I am presented with the message
> below. Since the test works the problem is not that the file can't be
> found - it is something else.
>
> I experience the same behavior in both QGIS2 and QGIS3. Is this the
> intended behavior? Is it by design that I am not able to use
> init-files or is this simply a bug? Is there a known workaround?
>
> [cid:image001.png at 01D38B96.E25BF930]
>
> There is no errors in the projection setup in the init-file, which
> the working test transformation proves. The init file in question can
> be seen here:
> https://github.com/NordicGeodesy/NordicTransformations/blob/master/resources/DK.
> I am aware that I can just add the definition of the CRS directly, or
> simply use epsg:25832. That is not the point. I have other more
> esoteric transformations in the same file that is more complicated to
> manage in QGIS itself.
>
> I appreciate any advice and guidance. If this indeed is a bug I will
> gladly help fix it.

QGIS (via GDAL) needs a working inverse projection to be defined. Many 
exotic projections defined in PROJ.4 fail therefore.

I guess the +proj=pipeline definitions spoil that requirement. I'm not 
even sure if those work in command-line GDAL.

Greetings,
Andre Joost