feat(python,rust,cli): add SQL round support

[alexander-beedie]

Add SQL support for ROUND function:

import polars as pl

df = pl.DataFrame( 
    {"n": [-8.499, 20.45678, 3.59901, 82.5001]} 
)
with pl.SQLContext( frame_data=df ) as ctx:
    print(
        ctx.execute("""
          SELECT 
            n,
            ROUND(n) AS n0,
            ROUND(n,1) AS n1,
            ROUND(n,2) AS n2, 
            ROUND(n,3) AS n3, 
          FROM
            frame_data
        """).collect()
    )
    
# shape: (4, 5)
# ┌──────────┬──────┬──────┬───────┬────────┐
# │ n        ┆ n0   ┆ n1   ┆ n2    ┆ n3     │
# │ ---      ┆ ---  ┆ ---  ┆ ---   ┆ ---    │
# │ f64      ┆ f64  ┆ f64  ┆ f64   ┆ f64    │
# ╞══════════╪══════╪══════╪═══════╪════════╡
# │ -8.499   ┆ -8.0 ┆ -8.5 ┆ -8.5  ┆ -8.499 │
# │ 20.45678 ┆ 20.0 ┆ 20.5 ┆ 20.46 ┆ 20.457 │
# │ 3.59901  ┆ 4.0  ┆ 3.6  ┆ 3.6   ┆ 3.599  │
# │ 82.5001  ┆ 83.0 ┆ 82.5 ┆ 82.5  ┆ 82.5   │
# └──────────┴──────┴──────┴───────┴────────┘

[ritchie46]

See panic comment.

[alexander-beedie]

See panic comment.

Yup; previous (post-comment) commit was just a rebase; will see about updating tonight.

[ritchie46]

@alexander-beedie got time for this one today? Otherwise, I will finish it so that we can include it in the release today.

[alexander-beedie]

@alexander-beedie got time for this one today? Otherwise, I will finish it so that we can include it in the release today.

Sorry - yes, if you could handle the polars_bail (or ensure) integration that would be great; I had a brief crack at it last night but managed to dig myself into a small hole; I'll work out what patently obvious step I was missing from your diff, heh 😅

[ritchie46]

Haha! Alright. :D