`pl.DataFrame.corr()` should offer an argument "include_column_names_as_column". about polars HOT 5 OPEN

@stinodego: Tihis reminds me of some similar recent additions, and I think there's a good ergonomic argument for adding it. Something like an optional label parameter perhaps? If None, no additional label col added (as now), if set to a string (eg: label="col") then add the label column with that name...🤔

from polars.

FYI: in the absence of a dedicated parameter, you can compose the labelled view like so -

df_corr = pl.DataFrame({
    "a": [1, 2, 3,],
    "b": [5, 6, 8,],
    "c": [-1, -2, -3]
}).corr()

df_corr.insert_column(0, pl.Series("col", df.columns))
# shape: (3, 4)
# ┌─────┬──────────┬───────────┬───────────┐
# │ col ┆ a        ┆ b         ┆ c         │
# │ --- ┆ ---      ┆ ---       ┆ ---       │
# │ str ┆ f64      ┆ f64       ┆ f64       │
# ╞═════╪══════════╪═══════════╪═══════════╡
# │ a   ┆ 1.0      ┆ 0.981981  ┆ -1.0      │
# │ b   ┆ 0.981981 ┆ 1.0       ┆ -0.981981 │
# │ c   ┆ -1.0     ┆ -0.981981 ┆ 1.0       │
# └─────┴──────────┴───────────┴───────────┘

from polars.

Yeah I'm aware. Just seemed nice and obvious to have it as an option, if not as the default. It's a similar situation to pl.DataFrame.describe() imo.

from polars.

I agree, it does seem like it should give that label column by default.

from polars.

@stinodego: Tihis reminds me of some similar recent additions, and I think there's a good ergonomic argument for adding it. Something like an optional label parameter perhaps? If None, no additional label col added (as now), if set to a string (eg: label="col") then add the label column with that name...🤔

That sounds good to me @alexander-beedie . I'd be happy to take this on as a PR :)

from polars.