Andy Teucher

Fantastic! Thanks @mdsumner. mapshaper does to post-checks and repairs intersections. From the [wiki](https://github.com/mbloch/mapshaper/wiki/Command-Reference#-simplify): > By default, mapshaper rolls back simplification along pairs of intersecting line segments by re-introducing removed points...

Oh your second point is very interesting! I'm not quite sure what you mean by doing data frame traversal first... right now it collects the information about the columns in...

> The scenario in mind is when the geometric conversion takes some time, and presumably succeeds - but then we fail on some data column - so could we fail...

I think it's going to require a bit more thought around dealing with objects that end up being aggregated/disaggregated (thinking especially `ms_dissolve()`, `ms_explode()`, `ms_simplify(explode = TRUE)`, but there may be...

@gaborcsardi another good point about debugging. And yes, it would need to be able to detect whether the next function in the pipe was from rmapshaper. It does seem like...

> Maybe as default `return_class = auto()` where `auto` will determine whether it is within a pipeline and whether it is the last such... That seems like a pretty nice...

That is intriguing! I'll probably keep it simple until I see how [sfr](/edzer/sfr) and [geojson](/ropenscilabs/geojson) play out, but thanks for the ideas!

Thanks @timelyportfolio! I'm really close to having something that people can kick the tires of. I'm trying to get all the methods implemented and tested for the five core functions...

Hi @bhaskarvk, thanks for the note. By convert, do mean the `format` output option listed [here](https://github.com/mbloch/mapshaper/wiki/Command-Reference#-o-output)? Unfortunately the `-o` options aren't currently exposed in the mapshaper command line API, so...

I think it should be possible, thanks for reporting this.