SNOW-2380579: TimestampType(s) should support precision argument

[ctc-andot]

What is the current behavior?

TimestampType only take timezone argument, not the precision argument that is available on the underlying snowflake type. https://docs.snowflake.com/en/sql-reference/data-types-datetime

What is the desired behavior?

They should take both, e.g. TimestampType(precision=6) to get a microsecond rather than nanosecond precision time.

How would this improve snowflake-snowpark-python?

Better mapping onto the full set of snowflake types. Additionally, with iceberg v2 only supporting microsecond precision, it is easy to accidentally lose precision when writing to iceberg table. This helps make that more explicit.

References, Other Background

[sfc-gh-sghosh]

Hello @ctc-andot ,

At present, Snowpark Python’s TimestampType currently only allows you to specify the timezone variant (NTZ, LTZ, TZ) for timestamp columns; it does not offer a way to set the fractional second precision (e.g., microseconds, nanoseconds) via its API. like Snowflake SQL.

As a workaround, please see if you can use the following workaround: create the data type with the required timestamp precision ( example for to use in Iceberg)

session.sql("CREATE OR REPLACE TABLE iceberg_ts_test(ts TIMESTAMP_LTZ(6))").collect(), so this way the desired precision will be maintained.

Will raise the feature request.

Regards,
Sujan