I sent Jukka and Andrea's comments to the developer. Here is the
response I received from him:
Thank you for you email and taking the time to actually
get in touch with GDAL maintainers and forward their replies to
me.
That's an interesting statement from them. I have been working
with shapefiles for the last 20 years and my personal
interpretation of M array being "optional" has always been
that in case no M dimensions are needed, some "no data"
values can be stored in the binary file.
For reference, on page 10 of the specs under "Measured
Shape Types in X,Y Space" paragraph it reads:
Shapes of this type have an additional coordinate M. Note that
"no data" value can be specified as a value for M (see Numeric
Types on page 2).
And on page 15 under "Shape Types in X,Y,Z Space"
paragraph, again:
Shapes of this type have an optional coordinate M. Note that
"no data" value can be specified as a value for M (see Numeric
Types on page 2).
On page 2, "Numeric Types" paragraph states:
[...] shapefiles support the concept of "no data" values, but
they are currently used only for measures.
Any floating point number smaller than –1038 is considered
by a shapefile reader to represent a "no data" value.
Thus, my interpretation has always been something like that in
case M array is not needed, I must fill it with "no data"
values.
Probably GDAL devs know more than me, but it seems odd that in
order to establish whether the M array is present or not in the
binary SHP file, one needs to check the SHX index file for the
binary address of the next record. Surely it could be, but it
feels cumbersome to me (not that it would be the only cumbersome
thing about shapefile specs :)) )
About the vast majority of GIS software being able to read those
shapefiles with missing M arrays, I have always thought they chose
to silently ignore the missing M array of the last record as a
favor to the user (yep, I know technically it's not just the last
record...).
Anyways, I am open to admit I have always been wrong with my
interpretation of the specs (maybe even relieved), it would be
great to have a confirmation from GDAL devs.
I think the source of the problem is confusion between "no data" and
"optional". I interpret "no data" as a null indicator, to fill in
required values but indicate that the value is meaningless, while
"optional" means the data need not appear at all. The developer is
not a native English speaker, which may have contributed to the
confusion.
_______________________________________________
gdal-dev mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/gdal-dev
