Well, in the end it is the pandoc markdown writes that is used to generate the tables, so you're on the right track with -t markdown+multiline_tables. However, unfortunately the logic is a bit more complicated, see https://github.com/jgm/pandoc/blob/master/src/Text/Pandoc/Writers/Markdown.hs#L444. Since all table-options are enabled by default, you have to turn OFF the kind of tables you don't want

, e.g. to get pipe-tables: pandoc --filter pandoc-placetable -t markdown-simple_tables

I agree that this is not optimal, but this behaviour should be changed in the markdown writer, not the filter...

On Sunday, May 29, 2016 at 5:59:49 PM UTC+2, kurt.p...-gM/Ye1E23mwN+BqQ9rBEUg@public.gmane.org wrote:

Am Sonntag, 29. Mai 2016 14:45:35 UTC+2 schrieb mb21:

@Kurt, pandoc-placetable works perfectly to convert csv to markdown table syntax (and any other output format where pandoc supports generating tables): 

pandoc --filter pandoc-placetable -t markdown

Sorry, I did not intend to mis-represent what pandoc-placetable currently can do and what it cannot.

I know it can convert to Markdown tables. But (AFAIU) it can generate only one type of table: simple_table.
However, with pandoc-csv2table I can generate simple_table, multiline_table, pipe_table and grid_table types — simply by adding it into the code block metadata: {.table header="yes" type="grid" ....}.

I tried to get the same thing with pandoc --filter pandoc-placetable -t markdown+multiline_tables, but it didn’t work.

(Maybe I’m missing something — then please tell me.)

​​