Neil B via QGIS-User <qgis-user@lists.osgeo.org> writes: > When you open a GPKG, meaning you are accessing data from the file, the > -wal file and its partner the -shm file are created in the same directory > as the GPKG file. The WAL file is created to implement atomic commit and > rollback. At the file system level, if you place the GPKG in read only mode > then there is never a concern about writing data to the database so the > -wal and -shm temp files are not created. The behaviour of the temporary > files has nothing to do whether the layer is put in edit mode or not, but > is a functionality of SQLite.
I thought that there was an improvement to gdal so that opening a gpkg layer read only would result in not creating temp files, and they would only arise when one edited the layer. I have always viewed situations where logical read operations lead to filesystem writes as buggy, intrinsically so, but they are very troublesome for distributed filesystems and version control. > You don't have the speed problem with shapefiles as they have no mechanism > to protect them from multiple users and no mechanism to ensure the > equivalent of an atomic commit. If multiple people make edits to the same > shapefile, the last one to save is the one that decides what is written, > assuming the shapefile doesn't get corrupted in the mean time. That's great perspective! _______________________________________________ QGIS-User mailing list QGIS-User@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-user Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
