B023 false alarms

Yep, this is one of those false alarms! I’d use a list-comprehension instead, which is both faster and IMO more readable:

def filter_values(values: list[list[int]], max_percentage: float):
    for value in values:
        filter_val = max(value) * max_percentage
        yield [x for x in values if x < filter_val]

Unfortunately it’s more difficult to avoid this false alarm than you might think: for example “lambdas are OK if they’re the first argument in a call to filter()” would miss bugs where the filter iterable (and therefore lambda) isn’t immediately evaluated.

Not the same usecase as described here originally, but I found a false alarm with the following code:

$ cat app.py
from app import db


for name in ['a', 'b']:
    def myfunc(name):
        qry = db.query(name=name)
        db.run(qry)

    myfunc(name=name)
    qry = db.query(description__contains=name)
    db.run(qry)

And when I run flake8:

$ venv/bin/flake8 app.py
app.py:7:16: B023 Function definition does not bind loop variable 'qry'.

Basically - if I have a variable defined within the loop - but the same variable name is also within the function. The local variable shouldn’t have any problem in terms of execution ?

But - the rule thinks it is an issue