[Octave-bug-tracker] [bug #61757] [octave forge](mapping) function "shaperead" Atttributes property/value document error

[Philip Nienhuis]

Update of bug #61757 (project octave):

             Assigned to:                    None => philipnienhuis         
                 Release:                   4.4.1 => other                  
        Operating System:               GNU/Linux => Any                    

    _______________________________________________________

Follow-up Comment #1:

(mapping pkg maintainer here)
Thanks for the bug report & the kind words.

You are obviously right in assuming that _in principle_ .shp file attribute
fields shouldn't also occur in the associated .dbf files. But ...
Below in shaperead.m there are some stanzas for prepending attributes present
as fields in the .shp file with underscores to distinguish them from the ones
in the .shp file. I wrote this so long ago that I cannot quickly recall now
why this was done; I suppose because at my work we do have (old or even
ancient; ArcView?) shapefiles where such fields occur in the .dbf's.

Don't ask me how that can be :-) I've even seen .dbf files with multiple
identical fieldnames - I suppose that is due to the attribute names in .dbf
files to be max 10 chars or so, implying truncation of too long fieldnames.

Now, looking at your bug description I tend to think the following:
* A bit of extra documentation for shaperead.m would do no harm.
* A robust solution needs to be implemented in dbfread.m (in the io package).
I'd suppose dbfread.m shouldn't choke on non-existent attribute (fieldname)
fields (i.e., ".shp" attribute names unconditionally prepended with
underscores). But maybe that check is already there, I'm unsure.

First I'll have to investigate a little and refresh my memory on what it was
all about. It may take a while before I have some results.

    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?61757>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/