`pct_change` is unable to find a column

[coastalwhite]

Checks

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of Polars.

Reproducible example

import polars as pl
df = pl.DataFrame({
    'a': [1, 2, 3],
})
df.select(
    x = pl.col.a,                                           # This works.
    b = pl.Series([1, 2, 5, 2]).pct_change(pl.col.a.min()), # This returns a col `a` not found error.
)

Log output

ColumnNotFoundError: a

Resolved plan until failure:

	---> FAILED HERE RESOLVING 'sink' <---
DF [""]; PROJECT */1 COLUMNS

Issue description

For some reason, the column a cannot be found here.

Expected behavior

It is found.

Installed versions

Replace this line with the output of pl.show_versions(). Leave the backticks in place.

[mcrumiller]

Are there any other cases where we can use an external column expression inside a pl.Series function? I have never seen that before. For example, I wouldn't expect this to work, and it doesn't:

df = pl.DataFrame({"a": [1, 2, 3]})
df.select(
    pl.Series([1, 2, 3]).append(pl.col("a"))
)

[coastalwhite]

Are there any other cases where we can use an external column expression inside a pl.Series function? I have never seen that before. For example, I wouldn't expect this to work, and it doesn't:

df = pl.DataFrame({"a": [1, 2, 3]})
df.select(
    pl.Series([1, 2, 3]).append(pl.col("a"))
)

Holy moly, you are right... I went through a few too many expressions in one day.