patch integrate in TIME does not update data_type

[dylanmikesell]

Description

Doing the CTEMPS tutorial and ran across a small bug when doing integration in time.
The data_units is properly updated after patch.integrate(), but the data_type is not updated.

Example

import dascore as dc
patch = dc.get_example_patch("example_event_2")
patch
patch_int = patch.integrate("time")
patch_int

Expected behavior

patch variable has attributes
data_type: strain_rate
data_units: ϵ / s

patch_int variable has attributes
data_type: strain_rate
data_units: ϵ

Versions

• OS: Ubuntu 24.04.2 LTS
• DASCore Version: 0.1.9
• Python Version: 3.12.3

[d-chambers]

Hi @dylanmikesell,

Thanks for reporting the issue.
Some background info: the data_type attribute was something we added a long time before units. At that time it served as an important way to know what kind of data a file contained. However, I suspect it may be obsolete/redundant now that we have units. I am a bit reluctant to remove it since some codes might depend on it.

Also, unlike with units, it is very hard to accurately update data_type through arbitrary operations. For example, integration/differentiation makes sense, but what about aggregations, or when dividing an array by a scalar with or without units? I don't think we can always meaningfully describe the data_type of many operations, so it becomes a can of worms to update it for some but not for others.

Maybe you have some thoughts on how to approach this?