map_dict(..., default=pl.first()) keeps the wrong column about polars HOT 3 CLOSED

can reproduce!

@stfdxv when you run your code you should get a warning that map_dict is deprecated

DeprecationWarning: map_dict is deprecated. It has been renamed to replace.

It works with the new replace method

pl.DataFrame({"a": [1], "b": [2], "c": [3]}).with_columns(
    c_mapped=pl.col("c").map_dict({0: 100}, default=pl.first()),
    c_replaced=pl.col("c").replace({0: 100}),
)

# shape: (1, 5)
# ┌─────┬─────┬─────┬──────────┬────────────┐
# │ a   ┆ b   ┆ c   ┆ c_mapped ┆ c_replaced │
# │ --- ┆ --- ┆ --- ┆ ---      ┆ ---        │
# │ i64 ┆ i64 ┆ i64 ┆ i64      ┆ i64        │
# ╞═════╪═════╪═════╪══════════╪════════════╡
# │ 1   ┆ 2   ┆ 3   ┆ 1        ┆ 3          │
# └─────┴─────┴─────┴──────────┴────────────┘

from polars.

@JulianCologne Yes, but see my next issue #15081.
In fact I had to dig through my code to get to the root problem. Only then I saw the description of map_dict() in my IDE saying it's deprecated.

from polars.

Since replace works correctly, I would like to refrain from changing the functionality of a deprecated method. So I'm closing this. Please use replace instead!

from polars.