CODEWITHSUNDEEP
dataframepython-polarspolars
Phil-ZXX
Phil-ZXX

Reputation: 3275

Rename all columns in dataframe to lowercase

Given a polars dataframe I want to rename all columns to their lowercase version. As per polars.Expr.name.to_lowercase we can do

import polars as pl
pl.DataFrame([{'CCY': 'EUR', 'Qty': 123},
              {'CCY': 'USD', 'Qty': 456}]).with_columns(pl.all().name.to_lowercase())

but this duplicates the data (as it keeps the original column names).

Conceptually, I am looking for something like

pl.DataFrame(...).rename({c: c.name.to_lowercase() for c in pl.all()})

But this doesn't work since pl.all() is not iterable.

Upvotes: 3

Views: 191

Answers (3)

mozway
mozway

Reputation: 262214

select rather than with_columns:

df.select(pl.all().name.to_lowercase())

Output:

┌─────┬─────┐
│ ccy ┆ qty │
│ --- ┆ --- │
│ str ┆ i64 │
╞═════╪═════╡
│ EUR ┆ 123 │
│ USD ┆ 456 │
└─────┴─────┘

Note that you could also use your envisioned approach with cs.expand_selector:

import polars.selectors as cs

df.rename({c: c.lower() for c in cs.expand_selector(df, cs.all())})

although in the case of all columns could be replaced by a simple:

df.rename(str.lower)

Upvotes: 4

roman
roman

Reputation: 117540

As said already, you can use select() instead of with_columns(). You can also use pl.DataFrame.rename() with function instead of dictionary:

pl.DataFrame([
    {'CCY': 'EUR', 'Qty': 123},
    {'CCY': 'USD', 'Qty': 456}
]).rename(str.lower)

shape: (2, 2)
┌─────┬─────┐
│ ccy ┆ qty │
│ --- ┆ --- │
│ str ┆ i64 │
╞═════╪═════╡
│ EUR ┆ 123 │
│ USD ┆ 456 │
└─────┴─────┘

You can also use pl.DataFrame.columns to get columns list, if you want to create dictionary:

df = pl.DataFrame([
    {'CCY': 'EUR', 'Qty': 123},
    {'CCY': 'USD', 'Qty': 456}
])

df.rename({c: c.lower() for c in df.columns})

shape: (2, 2)
┌─────┬─────┐
│ ccy ┆ qty │
│ --- ┆ --- │
│ str ┆ i64 │
╞═════╪═════╡
│ EUR ┆ 123 │
│ USD ┆ 456 │
└─────┴─────┘

Upvotes: 2

ouroboros1
ouroboros1

Reputation: 14369

Using df.rename + built-in str.lower:

df.rename(lambda col_name: col_name.lower())

Output:

shape: (2, 2)
┌─────┬─────┐
│ ccy ┆ qty │
│ --- ┆ --- │
│ str ┆ i64 │
╞═════╪═════╡
│ EUR ┆ 123 │
│ USD ┆ 456 │
└─────┴─────┘

Upvotes: 3

Related Questions