Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
FEAT 5460: Make DataFrameGroupBy.value counts more distributed (#3)
* FIX-modin-project#4154: add value_counts method for SeriesGroupBy and DataFrameGroupBy (modin-project#5453) Signed-off-by: Anatoly Myachev <anatoly.myachev@intel.com> * WIP GroupBy * Seemingly working w/ Ponder and Modin * Maintain proper value ordering for single group-by operations * linter updates from black on the files that have changed * When normalize is used, or the groupby was performed with as_index=False, default to pandas. With ponder this will result in a NotImplemented error currently. * Remove _to_pandas() by implementing sort_index on Series on the service side * | Object | call | Pushdown? | Compatibility | | --- | --- | --- | --- | | DataFrameGroupBy | value_counts() | Full | PASS | | DataFrameGroupBy | value_counts(ascending=True) |Full | PASS | | DataFrameGroupBy | value_counts(ascending=False) | Full |PASS | | DataFrameGroupBy | value_counts(sort=False) |Full | PASS | | DataFrameGroupBy | value_counts(sort=True) |Full | PASS | | DataFrameGroupBy | value_counts(normalize=False) |Full | PASS | | DataFrameGroupBy | value_counts(normalize=True) | | FAIL | | DataFrame | groupby(as_index=False) | | FAIL | | DataFrameGroupBy | value_counts(dropna=False) | Full | PASS | DataFrameGroupBy | value_counts(dropna=True) | |FAIL $\color{red}{\text{NOTE: For MultiIndex GroupBys the n+1 level index is ignored in the sorting.}}$ --------- Signed-off-by: Anatoly Myachev <anatoly.myachev@intel.com> Co-authored-by: Anatoly Myachev <anatoly.myachev@intel.com>
- Loading branch information