Community for the Daft project and all things distributed data

Distributed Data Community

Hey Folks,
I am unable to perform some operation in daft which is very crucial for my pipeline.
• In Pyspark we can do something =&gt; F.col("date_trans") - F.expr("interval 1 day") but here i am unable to find alternative
• In Pyspark we can subtract two dates =&gt; F.col("date_trans1") - F.col("date_trans2") but here i am unable to find alternative
Kindly suggest

Ah I think you’re looking for the Duration datatype! Cc <@U0664J23JMP> for any thoughts

Ah looks like we support the Duration datatype, but don’t yet support arithmetic using it :stuck_out_tongue:

Shouldn’t be too hard to build though! I’ll chat with the team and see if we can get something out this week

Hey <@U06TD184LKF>! We should be able to do arithmetic between `Timestamp` and `Duration` datatypes. E.g.

```import daft
import datetime

df = daft.from_pydict({"datetime": [datetime.datetime.now()], "timedelta": [datetime.timedelta(days=1)]})
df = df.with_column("diff", df["datetime"] - df["timedelta"])
df.show()```
We're going to work on arithmetic between `Date` and `Duration` next, as well as between `Date`/`Timestamp`

Here’s a tracking issue: <https://github.com/Eventual-Inc/Daft/issues/2098>