polars.Expr.struct.eval#

Expr.struct.eval(*exprs: IntoExpr | Iterable[IntoExpr], **named_exprs: IntoExpr) → Expr[source]#

Select fields of this struct, dropping the fields that are not selected.

This is similar to select on DataFrame. Use with_fields() to retain the fields that are not selected.

The expressions must be length-preserving; they are evaluated against the fields of the struct, which are addressed with polars.field().

engine:In MemoryStreamingDistributed

Parameters:

*exprs: Field(s) to select, specified as positional arguments. Accepts expression input. Strings are parsed as column names, other non-expression inputs are parsed as literals.
**named_exprs: Additional fields to select, specified as keyword arguments. The fields will be renamed to the keyword used.

See also

with_fields
field

Examples

>>> df = pl.DataFrame(
...     {
...         "coords": [{"x": 1, "y": 4}, {"x": 4, "y": 9}, {"x": 9, "y": 16}],
...         "multiply": [10, 2, 3],
...     }
... )
>>> df.with_columns(
...     pl.col("coords").struct.eval(
...         pl.field("x").sqrt(),
...         y_mul=pl.field("y") * pl.col("multiply"),
...     )
... )
shape: (3, 2)
┌───────────┬──────────┐
│ coords    ┆ multiply │
│ ---       ┆ ---      │
│ struct[2] ┆ i64      │
╞═══════════╪══════════╡
│ {1.0,40}  ┆ 10       │
│ {2.0,18}  ┆ 2        │
│ {3.0,48}  ┆ 3        │
└───────────┴──────────┘

Fields that are not selected are dropped:

>>> df.select(pl.col("coords").struct.eval(pl.field("y")))
shape: (3, 1)
┌───────────┐
│ coords    │
│ ---       │
│ struct[1] │
╞═══════════╡
│ {4}       │
│ {9}       │
│ {16}      │
└───────────┘