Hi @Joshua Pedrick! We don’t yet have a native implementation of

.datetime_range

in Daft (seems like it converts 2

timestamp

columns into a

list[timestamp]

column according to the provided interval) However, this is actually pretty easy to do via a UDF as an escape hatch:

import daft
import polars as pl
import datetime

@daft.udf(return_dtype=daft.DataType.list(daft.DataType.timestamp("us")))
def my_datetime_range(start: daft.Series, end: daft.Series, interval: str) -> pl.Series:
    polars_df = pl.DataFrame({"start": pl.Series(start.to_arrow()), "end": pl.Series(end.to_arrow())})
    polars_results = polars_df.select(range=pl.datetime_ranges(start="start", end="end", interval=interval))
    return daft.Series.from_arrow(polars_results["range"].to_arrow())

df = daft.from_pydict({"start": [datetime.date(2023, 1, 1), datetime.date(2023, 1, 2)], "end": [datetime.date(2023, 1, 5),
datetime.date(2023, 1, 7)]})
df = df.with_column("range", my_datetime_range(df["start"], df["end"], "1d"))

df.show()

We can make this a little cleaner too by supporting easier conversions between Daft

Series

and Polars

Series

. Let me know if that’s something you might want!

the polars function generates a series from given inputs

Ahh got it 🙂

Greater support for Polars conversion would be great just so I don't have to use my own helper functions!