Hi all and thanks for the discussion,
Even and I discovered that the .xls files are loading with correct
encoding in Linux but not on Windows which leads to the assumption that
freexl on Windows might not have been correctly compiled/linked against
iconv.
Something that maybe Jürgen could clarify?
Thanks,
Andreas
On 2021-07-15 23:14, Andrea Giudiceandrea wrote:
Il 15/07/2021 22:30, Nyall Dawson ha scritto:
I don't think that's completely correct -- looking at the freexl
documentation it seems that only some xls file versions are utf8, and
others have a codepage indicating the encoding which needs to be read
from the xls metadata:
Yes, of course BIFF5 (and previous versions), which is used until Excel
95, stores internally the strings with a codepage based encoding, while
BIFF8 (since Excel 98), stores them internally as UTF-16LE without the
need of codepages.
Anyway the documentation states that: "FreeXL is intended to be
strictly interoperable with SQLite and SpatiaLite, so any text string
has to be converted to UTF-8 encoding. GNU libiconv can easily handle
any required charset conversion. So we can simply fetch the appropriate
bytes, then call iconv() as appropriate, and we'll immediately get back
the corresponding UTF-8 encoded text string."
So, if I'm not mistaken, the decoding from the proper internal codec
and the encoding to UTF-8 is handled by the FreeXL library for both
BIFF2/BIFF5 and BIFF8.
I've also tested a BIFF5 XLS using ogrinfo, and the output is UTF-8
encoded, while internally
Regards.
Andrea
_______________________________________________
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
_______________________________________________
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
