Feature Request: support `df.group_by(...).agg(new_col: ...)` like `df.with_columns(new_col: ...)`
DeflateAwning opened this issue · 2 comments
DeflateAwning commented
It would be awesome if .agg()
supported the same semantics as .with_columns(...)
and .select(...)
. That is, I wish .agg
had support for naming new columns as kwargs to the method call.
Example which works with .with_columns
, but not with .agg
:
# works
df.with_columns(
volume_cumulative: Polars.col('volume').cum_sum,
)
# doesn't work
df.group_by('date').agg(
volume_cumulative: Polars.col('volume').sum,
)
ankane commented
Hi @DeflateAwning, added in the commit above.
DeflateAwning commented
Amazing, thanks!