replace_strict with default values

[edenian-prince]

I'm trying to map values from a dict. I have a column of values in my dataframe, and want something like if apples is a value in my dataframe column and it is in the dict, then return the value mapped to apples, which is red. I need a default value where if a value in my dataframe column is not in the dict, then it would assign default.

it looks like default is not in narwhals currently. is there a simple way to use replace_strict and assign default values?

in polars it's pretty simple:

# Create simple test DataFrame
df = pl.DataFrame({
    "fruit": ["apple", "banana", "cherry", "durian"]
})

# Mapping (note: "durian" is missing!)
mapping = {
    "apple": "red",
    "banana": "yellow",
    "cherry": "red"
}

df.with_columns(
   pl.col("fruit").replace_strict(mapping,default='default_string')
)

Here's an example with narwhals:

import polars as pl
import narwhals as nw

# Create simple test DataFrame
df = pl.DataFrame({
    "fruit": ["apple", "banana", "cherry", "durian"]
})

# Convert to Narwhals
data = nw.from_native(df)

# Mapping (note: "durian" is missing!)
mapping = {
    "apple": "red",
    "banana": "yellow",
    "cherry": "red"
}

# Attempt to apply replace_strict
# Expected: unmatched values fallback to default
# Actual: InvalidOperationError, and default= is NOT supported
keys = list(mapping.keys())
values = list(mapping.values())

data.with_columns(
    nw.col("fruit")
    .replace_strict(keys, values)  # ⚠️ No `default` param allowed
    .alias("color")
)

And I get an error like this, even though there is no default parameter:

InvalidOperationError: incomplete mapping specified for `replace_strict`

Hint: Pass a `default` value to set unmapped values.
Cell In[22], line 7
      4 keys = list(mapping.keys())
      5 values = list(mapping.values())
----> 7 data.with_columns(
      8     nw.col("fruit")
      9     .replace_strict(keys, values)  # ⚠️ No `default` param allowed
     10     .alias("color")
     11 )

there's gotta be a simple way to do this? i tried .with/then statements but face other issues, like it's trying to assign the whole list of values to a specific row

TypeError: cannot create expression literal for value of type Expr.

Hint: Pass `allow_object=True` to accept any value and create a literal of type Object.
(
    data
    .with_columns(
        nw.when(nw.col("lab").is_in(keys))
    .then(nw.col("fruit").replace_strict(keys, values))  # not strict!
    .otherwise(pl.lit("default_string"))
    )
)

[MarcoGorelli]

thanks @edenian-prince for the report

yeah we should have a default argument there

[MarcoGorelli]

I've made this into a feature request: #3209