[PyQGIS] as_numpy(use_masking=True) doesn't work well

[kadyb]

What is the bug or the crash?

I noticed two problems with as_numpy(use_masking=True):

1. In case of QgsRasterLayer it doesn't return masked array.
2. In case of QgsRasterBlock it doesn't use the correct fill value (should be 9999.0 in the example below).

Steps to reproduce the issue

Download this sample dataset from the repository: https://github.com/kadyb/adg2024/blob/main/dane/DEM.tif
Run the code below:

from qgis.core import QgsRasterLayer
import numpy.ma as ma

raster = QgsRasterLayer("DEM.tif", "DEM")
print(raster.isValid())
#> True
print(raster.dataProvider().sourceNoDataValue(1))
#> 9999.0

## QgsRasterLayer
mat = raster.as_numpy(use_masking = True)
print(ma.isMaskedArray(mat))
#> False

## QgsRasterBlock
block = raster.dataProvider().block(1, raster.extent(), raster.width(), raster.height())
print(block.isValid())
#> True
mat = block.as_numpy(use_masking = True)
print(ma.isMaskedArray(mat))
#> True
print(mat.fill_value)
#> 0.0   # should be 9999.0

Versions

QGIS 3.40.0, Windows 10

Supported QGIS version

• I'm running a supported QGIS version according to the roadmap.

New profile

• I tried with a new QGIS profile

Additional context

No response

[agiudiceandrea]

Gentle ping @merydian.