Coalesce gives different output with pl.col

[braaannigan]

Polars version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of Polars.

Issue description

I get incorrect output from pl.coalesce if I pass the columns inside pl.col instead of just as a list of strings - it's only using the first column in the expression.

Reproducible example

import polars as pl
df = pl.DataFrame(
    {
        "a":[10,None,22,None],
        "b":[8,12,19,1],        
        "c":[5,1,19,None],
    }
)
(
    df
    .with_column(
        pl.coalesce(pl.col(["c","b","a"])).alias("d")
    )
)
shape: (4, 4)
┌──────┬─────┬──────┬──────┐
│ a    ┆ b   ┆ c    ┆ d    │
│ ---  ┆ --- ┆ ---  ┆ ---  │
│ i64  ┆ i64 ┆ i64  ┆ i64  │
╞══════╪═════╪══════╪══════╡
│ 10   ┆ 8   ┆ 5    ┆ 10   │
├╌╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌┼╌╌╌╌╌╌┤
│ null ┆ 12  ┆ 1    ┆ null │
├╌╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌┼╌╌╌╌╌╌┤
│ 22   ┆ 19  ┆ 19   ┆ 22   │
├╌╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌┼╌╌╌╌╌╌┤
│ null ┆ 1   ┆ null ┆ null │
└──────┴─────┴──────┴──────┘

Expected behavior

Expect to get the same behaviour as this

(
    df
    .with_column(
        pl.coalesce(["c","b","a"]).alias("d")
    )
)
shape: (4, 4)
┌──────┬─────┬──────┬─────┐
│ a    ┆ b   ┆ c    ┆ d   │
│ ---  ┆ --- ┆ ---  ┆ --- │
│ i64  ┆ i64 ┆ i64  ┆ i64 │
╞══════╪═════╪══════╪═════╡
│ 10   ┆ 8   ┆ 5    ┆ 5   │
├╌╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌┼╌╌╌╌╌┤
│ null ┆ 12  ┆ 1    ┆ 1   │
├╌╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌┼╌╌╌╌╌┤
│ 22   ┆ 19  ┆ 19   ┆ 19  │
├╌╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌┼╌╌╌╌╌┤
│ null ┆ 1   ┆ null ┆ 1   │
└──────┴─────┴──────┴─────┘

Installed versions

---Version info---
Polars: 0.14.27
Index type: UInt32
Platform: Linux-5.10.104-linuxkit-x86_64-with-glibc2.28
Python: 3.10.1 (main, Dec 21 2021, 09:50:13) [GCC 8.3.0]
---Optional dependencies---
pyarrow: 10.0.0
pandas: 1.5.1
numpy: 1.23.4
fsspec: <not installed>
connectorx: 0.3.1
xlsx2csv: 0.8
matplotlib: <not installed>