feat: Increase verbosity of duplicate column error message

[mcrumiller]

Resolves #14458

I notice that none of our rust errors are multi-line, is this okay here?

wrong button

[mcrumiller]

I cannot replicate the hypothesis error. It appears it was failing in this case:

import polars as pl
from datetime import date, datetime

pl.DataFrame({
    "col0": pl.Series([0, 0], dtype=pl.UInt16),
    "col1": pl.Series([0, 0], dtype=pl.UInt16),
    "col2": pl.Series([datetime(2000, 1, 1), datetime(2000, 1, 1)], dtype=pl.Datetime("ns")),
    "col3": pl.Series([0, 0], dtype=pl.UInt16),
    "col4": pl.Series([0, 0], dtype=pl.UInt8),
    "col5": pl.Series([None, None], dtype=pl.Date),
    "col6": pl.Series([None, None], dtype=pl.Categorical),
}).null_count()

shape: (1, 7)
┌──────┬──────┬──────┬──────┬──────┬──────┬──────┐
│ col0 ┆ col1 ┆ col2 ┆ col3 ┆ col4 ┆ col5 ┆ col6 │
│ ---  ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ ---  ┆ ---  │
│ u32  ┆ u32  ┆ u32  ┆ u32  ┆ u32  ┆ u32  ┆ u32  │
╞══════╪══════╪══════╪══════╪══════╪══════╪══════╡
│ 0    ┆ 0    ┆ 0    ┆ 0    ┆ 0    ┆ 2    ┆ 2    │
└──────┴──────┴──────┴──────┴──────┴──────┴──────┘

This should pass the assertion test, no clue why it's reporting 2. I installed the same hypothesis version and set the replication decorator, but it fails to run on my machine.

[alexander-beedie]

I cannot replicate the hypothesis error.

It's this: #11910

[mcrumiller]

Thanks Alex. Once that gets resolved I'll pull those changes in here.

[mcrumiller]

@ritchie46 this one is now good to go.

[ritchie46]

This is still to broad IMO. I think this should be here:

polars/crates/polars-lazy/src/physical_plan/executors/projection_utils.rs

Line 239 in a28b922

polars_ensure!(names.insert(name), duplicate = name);

[stinodego]

@mcrumiller I opened a new issue specifically for the improvement in this PR. Could you address Ritchie's comment and finish this improvement?

[mcrumiller]

@stinodego I think I accomplished what he suggested, let me know if it still needs work and/or you have any suggestions.

Edit: yikes hang on, there's some bad formatting.

[mcrumiller]

Edit: fixed. Got some failures due to the new CI issue.

[mcrumiller]

Should we add newlines for our long error messages, or let the terminal wrap them?

>>> import polars as pl
>>> df = pl.DataFrame({"a": [1], "b": [1]})
>>> df.with_columns(pl.all().alias("a"))
polars.exceptions.ComputeError: the name: 'a' passed to `LazyFrame.with_columns` is duplicate

It's possible that multiple expressions are returning the same default column name. If this is the case, try renaming the columns with `.alias("new_name")` to avoid duplicate column names.
>>> df.lazy().with_columns(pl.all().alias("a")).collect()
polars.exceptions.ComputeError: the name: 'a' passed to `LazyFrame.with_columns` is duplicate

It's possible that multiple expressions are returning the same default column name. If this is the case, try renaming the columns with `.alias("new_name")` to avoid duplicate column names.

[stinodego]

Should we add newlines for our long error messages, or let the terminal wrap them?

Terminal should wrap them.

[mcrumiller]

I'll rebase to fix the CI tests.