polars.Series.shrink_dtype

Series.shrink_dtype() → Series

Shrink numeric columns to the minimal required datatype.

Shrink to the dtype needed to fit the extrema of this [Series]. This can be used to reduce memory pressure.

Examples

>>> s = pl.Series("a", [1, 2, 3, 4, 5, 6])
>>> s
shape: (6,)
Series: 'a' [i64]
[
    1
    2
    3
    4
    5
    6
]
>>> s.shrink_dtype()
shape: (6,)
Series: 'a' [i8]
[
    1
    2
    3
    4
    5
    6
]