Cast multiple columns with Unix epoch to Datetime

I have a dataframe with multiple columns containing unix epochs.
In my example I only use 2 of 13 columns I have. I’d like to cast all those columns to a datetime with UTC timezone in a single call to with_columns().

Polars provides the user with pl.from_epoch. However, I didn’t find a way to apply it to multiple columns as once.

df = pl.DataFrame({
    "id": [1, 2, 3, 4, 5],
    "start_date": [1566637530, 1561372720, 1561374780, 1558714718, 1558715044],
    "end_date": [1566628686, 1561358079, 1561358135, 1556188225, 1558427697],
    "cancelable": [True, True, False, False, True],
})

but I need:

So far, my code looks as follows.

columns_epoch_to_timestamp: list[str] = [
    "start_date",
    "end_date",
]

df = df.with_columns(pl.col(*columns_epoch_to_timestamp). )

>Solution :

You can pass pl.col("start_date", "end_date") to pl.from_epoch()

df.with_columns(pl.from_epoch(pl.col("start_date", "end_date")))

shape: (5, 4)
┌─────┬─────────────────────┬─────────────────────┬────────────┐
│ id  ┆ start_date          ┆ end_date            ┆ cancelable │
│ --- ┆ ---                 ┆ ---                 ┆ ---        │
│ i64 ┆ datetime[μs]        ┆ datetime[μs]        ┆ bool       │
╞═════╪═════════════════════╪═════════════════════╪════════════╡
│ 1   ┆ 2019-08-24 09:05:30 ┆ 2019-08-24 06:38:06 ┆ true       │
│ 2   ┆ 2019-06-24 10:38:40 ┆ 2019-06-24 06:34:39 ┆ true       │
│ 3   ┆ 2019-06-24 11:13:00 ┆ 2019-06-24 06:35:35 ┆ false      │
│ 4   ┆ 2019-05-24 16:18:38 ┆ 2019-04-25 10:30:25 ┆ false      │
│ 5   ┆ 2019-05-24 16:24:04 ┆ 2019-05-21 08:34:57 ┆ true       │
└─────┴─────────────────────┴─────────────────────┴────────────┘