Group by¶

Group a table by column values and aggregate each group. These methods are available on pykx.Table and pykx.KeyedTable objects.

The aggregations sum, prod, min, max, mean, median, count, std, sem, skew, kurt, idxmax, idxmin, any and all can be called on the result of groupby. They are described on the metadata and aggregations page.

groupby ¶

groupby(
    by=None,
    axis=0,
    level=None,
    as_index=True,
    sort=True,
    group_keys=True,
    observed=False,
    dropna=True,
)

Group the table by the values of one or more columns.

Parameters:

Name	Type	Description	Default
`by`	`str, int or list`	The name of the column, or a list of names, to group by.	`None`
`axis`	`int`	Not yet implemented, must be 0.	`0`
`level`	`str, int or list`	The name or number of the key column, or a list of them, to group by. A key number can only be used on a `KeyedTable`. `by` and `level` cannot both be given.	`None`
`as_index`	`bool`	Return the groups as the key of the result.	`True`
`sort`	`bool`	Sort the result by the group keys.	`True`
`group_keys`	`bool`	Not yet implemented, must be `True`.	`True`
`observed`	`bool`	Not yet implemented, must be `False`.	`False`
`dropna`	`bool`	Drop the groups where any key is null.	`True`

Returns:

Type	Description
`GroupbyTable`	A `GroupbyTable`, on which aggregations such as `sum` can be called.

Examples:

>>> tab = kx.q('([] k:`a`b`a; v:1 2 3)')
>>> tab.groupby('k').sum()
pykx.KeyedTable(pykx.q('
k| v
-| -
a| 4
b| 2
'))
>>> tab.groupby('k', as_index=False).sum()
pykx.KeyedTable(pykx.q('
idx| k v
---| ---
0  | a 4
1  | b 2
'))