Specify column types by name for `df_from_csv()`

[Columbus240]

How to specify the column types by name when using df_from_csv()? The current documentation (via #140, #189) only gives an example with positional arguments, which is particularly unwieldy with 10+ columns.

[krlmlr]

Thanks for the suggestion. Documentation updates are incoming.

The stable interface is called read_csv_duckdb() .

[krlmlr]

Need to review the behavior, tests, and documentation. Some of that may be wishful thinking.

[fmfrisch]

Hi, yes it seems that passing the options=list(types=list(c(ColA = "VARCHAR", ColB = "BIGINT"))) to read_csv_duckdb ignores the names and continues to use the position, so column 1 will be read as VARCHAR and column 2 as BIGINT. ColA and ColB will be unaltered unless they are in the first and second position.