You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
PIXELTYPE: This may appear on a GDT_Byte band (or the corresponding dataset) and have the value SIGNEDBYTE to indicate the unsigned byte values between 128 and 255 should be interpreted as being values between -128 and -1 for applications that recognise the SIGNEDBYTE type. Starting with GDAL 3.7, this metadata item is no longer used, as the Int8 data type is used for signed bytes.
It is not immediately clear to me if that means that the SIGNEDBYTE flag will be ignored entirely or if we should expect pygeoprocessing to handle both cases. We should investigate this behavior and make a decision on how to handle this. Until we do, signed integer raster operations in GDAL 3.7+ may have unexpected behavior.
The text was updated successfully, but these errors were encountered:
Can confirm that GDAL 3.7+ has changed behavior and causes tests to fail. GDT_Int8, added in 3.7, has a constant representation of 14. This is attempting to be mapped to _GDAL_TYPE_TO_NUMPY_LOOKUP, which does not include it.
Per the GDAL docs:
It is not immediately clear to me if that means that the SIGNEDBYTE flag will be ignored entirely or if we should expect pygeoprocessing to handle both cases. We should investigate this behavior and make a decision on how to handle this. Until we do, signed integer raster operations in GDAL 3.7+ may have unexpected behavior.
The text was updated successfully, but these errors were encountered: