Problems using select() with pandas 3.0.0

[snow73]

What happened?

Hi,

I have noticed an issue when using pandas 3.0.0 and using select() on a fieldset. It was no longer possible to select fields based on endStep (returned fieldset was None) while selecting using shortName and step worked fine. If I explicitly work with pandas 2.3.3 this problem does not arise.

Maybe this actually applies to all keys which are not part of the standard set (e.g. listed with ls()), but I didn't test that yet.

Best regards
Dennis

What are the steps to reproduce the bug?

new = mv.regrid(data=f.select(shortName=parameter, endStep=step), grid=[2, 2])

pip install metview pandas==2.3.3
--> works
pip install metview pandas==3.0.0
--> doesn't work

Version

v1.16.2

Platform (OS and architecture)

Linux b1548b5f5c57 6.12.67-linuxkit #1 SMP Sun Jan 25 02:26:28 UTC 2026 aarch64 aarch64 aarch64 GNU/Linux

Relevant log output

Accompanying data

No response

Organisation

MeteoIQ GmbH

[iainrussell]

Hi @snow73,
Thanks for reporting this! I can reproduce it. It seems that pandas 3.0 is stricter in type-checking. We have a set of pre-defined keys with their preferred types (int/str/etc) - step is in there, but endStep is not. The quick workaround is to change your query to put quotes around your endStep value to make it a string: step="90". If you think this will be a problem, we can look at making endStep an int by default.

[snow73]

Hi @iainrussell, no worries, I have downgraded to pandas 2.3.3 for the time being. Thanks for pointing out the workaround!