4

Currently when I try to retrieve date from a polars datetime column, I have to write sth. similar to:

df = pl.DataFrame({
    'time': [dt.datetime.now()]
})


df = df.select([
    pl.col("*"),
    pl.col("time").apply(lambda x: x.date()).alias("date")
])

Is there a different way, something closer to:

pl.col("time").dt.date().alias("date")
Alex
  • 439
  • 5
  • 16

2 Answers2

8

You can cast a Datetime column to a Date column:


import datetime
import polars as pl

df = pl.DataFrame({
    'time': [datetime.datetime.now()]
})

df.with_column(
    pl.col("time").cast(pl.Date)
)

shape: (1, 1)
┌────────────┐
│ time       │
│ ---        │
│ date       │
╞════════════╡
│ 2022-08-02 │
└────────────┘
ritchie46
  • 10,405
  • 1
  • 24
  • 43
2

Since polars version 0.16.15 (released just a few days ago). The code in your question is actually valid.

pl.col("time").dt.date().alias("date")

This is the diff that made it possible. And it works both on pl.Series and pl.Expr (used in lazy data frames).

Jan Kislinger
  • 1,441
  • 14
  • 26