use-fstring-for-concatenation can change behaviour in pandas

[Hellebore]

Issue description or question

I came across a suggested reformatting with Pandas code that changed the result. The suggestion is "Sourcery - Use f-string instead of string concatenation" on code that looks like this:
df["game_and_track"] = df["game"] + "--" + df["track_title"]
The suggested refactoring changed the operation from broadcasting the string to the shape of the other columns and vectorizing the operation to joining the entirety of each column to the string n times (n=the number of rows).

Sourcery Version

0.10.2

Code editor or IDE name and version

All

OS name and version

All

[Hellebore]

This is difficult for us to deal with - for the moment am going to disable this refactoring when pandas has been imported.