implicit casting for dt.datetime

[emanuel-schmid]

Changes proposed in this PR:

• implicit casting to integer when calling dt.datetime()

This PR fixes #96

PR Author Checklist

• Read the Contribution Guide
• Correct target branch selected (if unsure, select develop)
• Descriptive pull request title added
• Source branch up-to-date with target branch
• Documentation updated
• Tests updated
• Tests passing
• No new linter issues
• Changelog updated

PR Reviewer Checklist

• Read the Contribution Guide
• CLIMADA Reviewer Checklist passed
• Tests passing
• No new linter issues

[tovogt]

Why is this necessary? The used dt.year, dt.month and dt.day accessors should return integer types already, right? https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Series.dt.month.html

[emanuel-schmid]

@tovogt right! But that will eventually have changed by python 3.11.

(However - since python 3.11 also seems to inhibit the lint stage of the jenkins pipeline I'm going to re-pin it to python 3.9.)

[tovogt]

Ah, now I get it. The problem is not that the dtype is wrong, but the problem is that the expression track.time.dt.year[0] still returns a DataArray object with zero dimensions (i.e., a scalar). In this case, I suggest to use the expression track.time.dt.year.values[0] instead (as is already done in climada.hazard.trop_cyclone in CLIMADA-project/climada_python@7ba187b).

[tovogt]

Suggestions to solve this issue similar to the code in climada.hazard.trop_cyclone (CLIMADA-project/climada_python@7ba187b)

[emanuel-schmid]

🙌 @tovogt, perfect!