Read GeoJSON

[jayvdb]

https://github.com/wireservice/csvkit does a really nice job of reading GeoJSON into a CSV with their geojs.py, moving all of the properties into columns.

pyexcel-text has JSON support, which is a bit odd as this is mostly a tabulate-based plugin.

I suspect that geojson support should be in a separate plugin, and maybe it would also support topojson.

[jayvdb]

c.f. #30 for ndjson and other stream json derivative formats.

[chfw]

on top of my head, I would imagine a pyexcel-schema component to be created first to do semantic parsing. And then based on that, pyexcel-geojson could be written to specify persistent format(or layout) using tabular data structure.

[chfw]

being json format, pyexcel-geojson could be put under pyexcel-json as future version so as to minimze the number of package.

[chfw]

Could you please also shred the light on what you could use the geojson reading capability for? What I could see is to do transcoding from geojeson to csv, xls, etc.

[jayvdb]

My use case is BesutKode/uni-task-2#12 ; I want to sort the GeoJSON, so that this file format is git-friendly during PRs.

I will also want to do filtering of unneeded features, later.

I dont need a schema; I am quite happy with the complex objects being blobs in the CSV. I've made quite good progress in round-tripping the GeoJSON with csvkit : wireservice/csvkit#868 - only a few remaining issues.

[chfw]

I had a read on the geojs.py. I don't think dumping the geometry object as a json string is useful at all. In a excel software, that string is too long to fit in a cell, hence hard to read. So I think the geometry field is mostly ignored. My question is: if the other fields in geojson are of interest, why not skipping the blob? The only reason to have it is to re-construct the geo json back but csv is not a good format. Multi polygon has 4 dimensional array. I am still wondering how I put it in csv format if I am not doing simple json dump.