I setup my filter in ickc/pantable: CSV Tables in Markdown: Pandoc Filter for CSV Tables and called it pantable. I kind of want to break to pattern of pandoc-... since it is too cumbersome and Python doesn’t like -. In a sense it also emphasize pantable is a subset of pandoc, all about tables. For the meanwhile it only support CSV input. (But I’m thinking about adding .xlsx (seems trivial) and .html (less trivial) input. The last one might seems redundant, but basically the idea is to use HTML to type tables when it is too complicated, but rather than being a raw HTML I want it be real pandoc Table element for other outputs.)

I wrote extensive tests to make sure it works in cornering case. One thing that could be controversial is error handling. Basically for the metadata part, whenever it is invalid (e.g. header: first row while I expect boolean or yes/no), it will be overridden with sensible defaults. I think it is kind of like pandoc’s spirit that “any random string is valid markdown”, so I try to suppress and handle away all the errors. But I’m not sure if it will makes the error too difficult to be spotted.

Please do give feedback if you have any. I will put it into real world use in my project heavily in the coming month.

Thanks.