I think that the zeros or spaces in a DBF files are used only as padding
values, since the fields have fixed size. But I think this is an
format-specific details which should not be visible from the user,
especially if we want the DataStore API to be format-neutral. We do not
yet have a ShapeFile writer, but I think it would be writer's job to
re-insert padding values as needed.
For example one could read data from a text file and write them in a
ShapeFile. Consequently, I think that the Java objects after the reading
process should be format-neutral as much as possible, and the writer
should take care of format-specific details like padding values.
Martin
Le 30/09/14 20:13, Marc Le Bihan a écrit :
> Thanks a lot.
>
> I have to investigate about these zeros. Are they expected in a DBF
> file ? Or have we only come upon an odd shapefile ?
> I will try to load few shapefile from everywhere to see how the fields
> names are usually declared.
> Because I don't know if Apache SIS is already able to save a
> Shapefile, but maybe if we load a shapefile "with terminal zeroes", we
> are expected to save this shapefile with these zeros too ?
>
> Marc.
