Why? Implementation-wise, it's far better to have a pure plpgsql function than a Python UDF if you have a choice between the two. A Python UDF would be useful if you need to do more Python stuff in the db in general.
I think often times folks who come to postgres often don't know PL/pgsql (as I also didn't) and prefer to user their language of choice. I truly think if our postgres dev environments were better, we'd find more people writing code this way, so I can empathize with why people would prefer python or JS in the db, but at this stage I can't get myself to use those extensions.
FWIW, at one point, AWS didn't support a lot of these (sometimes unsafe) language extensions. So going pure PL/pgsql was the only true path forward. I think now for AWS users, they do support plv8 now, but as you pointed out somewhere in this thread, they are much slower... also I remember plv8 specifically had memory leaks.
Is it performance related? Not requiring another language to be installed?
It’s a bit of a weird case where I needed to reverse arrays inside a big json blob (millions of big json blobs). Trying to explode the structures out and rebuild them was a bit of a no-go. Instead I did a nasty little regex replace on the json string:
re.sub('"key": \[([^\\]]+)\]', lambda match: '"key": [' + ', '.join(reversed(match.group(1).split(', '))) + "]", els)
Since then I've always installed the python extension whenever I set up a new postgres, just in case.