Geopandas: Implement force_3d

[chay0112]

Use case:

• Add support for GeoSeries.force_3d

Description:

• Implements GeoSeries.force_3d, which converts each geometry in the series to a 3D geometry by adding a Z coordinate. If a geometry is already 3D, it is returned unchanged; otherwise, a default Z value (typically 0.0) is added to all coordinates. This ensures all geometries in the series have three dimensions.

Related Issue:

• Links to epic [EPIC] Implement More Geopandas Functions #2230

[chay0112]

I am working on it.

[chay0112]

@petern48
QQ: In GeoPandas, the force_3d function accepts both scalar and array-like values for the Z coordinate, allowing users to specify a different Z value for each geometry. However, since Sedona uses Spark SQL under the hood, does it make sense to support array-like Z values in this context? Specifically, can Sedona's Spark SQL backend handle per-row Z values, or should we only support scalar Z values for all geometries in the series?
Atleast here, I don't see the array like example use-case. Correct me if I am wrong.
No rush ! Please reply at your own convenience.

[petern48]

Geopandas: z : float | array_like (default 0)
Sedona: ST_Force3D(geometry: Geometry, zValue: Double)
PostGIS (what Sedona generally tries to follow): geometry ST_Force3D(geometry geomA, float Zvalue = 0.0);

Yeah, support array arguments isn't too common for SQL, and it's honestly slower for Sedona, since it requires a distributed shuffle + join (_row_wise_operation()). But we do have a convenient way to do it. Model it off the implementation of .segmentize(). I noticed a few bugs in .segmentize(), so I submitted this PR to fix a few things. Here's the link to the .segmentize() implementation in that PR. You'll likely need to copy over this change for lists to work properly or wait for that PR to get merged.

Great Question!

[petern48]

Alright, it's been merged. Make sure to pull main

[chay0112]

Alright, it's been merged. Make sure to pull main

Thanks Peter ! Will do.

[chay0112]

Alright, it's been merged. Make sure to pull main

@petern48 Sorry for taking a while for this PR, was busy with my work. The PR is ready for review now. Thanks for your patience.