enh: A way to check if a data type is Timestamp

[MarcoGorelli]

Is your feature request related to a problem?

Currently, it looks like daft_dtype == daft.DataType.timestamp('us', None) can check if daft_dtype is a Timestamp with microsecond precision and no time zone

But, how can I check if daft_dtype is any timestamp (regardless of time unit / time zone)?

Describe the solution you'd like

is_timestamp or something like that

Describe alternatives you've considered

No response

Additional Context

narwhals integration

Would you like to implement a fix?

No

[rchowell]

Would you like to do this in python code, or with a daft expression?

For example, do you want to do something like (1) or (2)

# (1) create a new column in the df with true/false if `my_col` is a timestamp
df.with_column("is_timestamp", col("my_col").is_timestamp())

# (2) or something like check the dtype in a schema?
df.schema.field.dtype.is_timestamp()

• (1) is interesting because this is a type check expression which unlocks some cool use cases and is applicable to SQL
• (2) is like a python isinstance thing so this would not work using Daft SQL

[MarcoGorelli]

thanks! the second

[rchowell]

Perfect! Cory has a PR out to address (2) 👍

#4034