polars
polars copied to clipboard
Allow duplicate columns in sub-parts of the graph if the final result is valid
Description
I am in a situation like so:
a = pl.LazyFrame({"a": [1, 2], "x": [3, 4]})
b = pl.LazyFrame({"b": [3, 4], "x": [5, 6]})
c = pl.concat([a, b], how='horizontal', strict=True)
d = c.select("a")
I have multiple intermediate LazyFrames with a duplicate column. I am concatenating these lazyframes, but at the end I'm only selecting from one of them. This fails because the intermediate LazyFrame is invalid (duplicate column), but column projection rules should mean that it doesn't matter.
Should this be something that is allowed?
I don't think I want to open this can of worms.