sqlp selects wrong data when multiple tables have the same named column

[urkle]

Describe the bug
qsv sqlp selects the wrong "table" when multiple CSV files have the same-named column.

To Reproduce
Steps to reproduce the behavior:

1. create CSV file "one.csv" with contents

id,data
1,open

2. create CSV file "two.csv" with contents

id,data
1,closed

3. run `qsv sqlp one.csv two.csv 'SELECT _t_1.id, _t_2.data FROM _t_1 JOIN _t_2 ON _t_1.id = _t_2.id'
   The wrong output is produced

id,data
1,open

Expected behavior
The output

id,data
1,closed

Desktop (please complete the following information):

• OS: macOS 12.7.4
• qsv Version: qsv 0.127.0-standard-apply;fetch;foreach;geocode;Luau 0.622;to;polars-0.39.2;self_update-16-16;51.20 GiB-1.43 GiB-22.83 GiB-64.00 GiB (x86_64-apple-darwin compiled with Rust 1.77.2) prebuilt

[urkle]

This looks to be an issue with polars.

I installed polars-cli 0.7.0 and ran this query.

select a.id, a.data as d1, b.data as d2 from read_csv('one.csv') as a join read_csv('two.csv') as b ON a.id = b.id;

which yields

┌─────┬──────┬──────┐
│ id  ┆ d1   ┆ d2   │
│ --- ┆ ---  ┆ ---  │
│ i64 ┆ str  ┆ str  │
╞═════╪══════╪══════╡
│ 1   ┆ open ┆ open │
└─────┴──────┴──────┘

[jqnatividad]

Thanks @urkle for the report and checking the query on polars-cli as well.

As you surmised, sqlp is really a convenience wrapper to tap the power of Polars for CSVs from the command-line.

Given polars' development tempo, if you report the issue, I wouldn't be surprised if the fix makes it into the next Rust Polars release.

Also, it's been about a month since polars 0.39.2 has been released, and there are tons of improvements I'm looking forward to...

pola-rs/polars@rs-0.39.2...main

[urkle]

@jqnatividad I am working on reporting w/ polars right now.

[jqnatividad]

pola-rs/polars#16255 was closed as the problem was originally reported in pola-rs/polars#15929, which is still awaiting resolution.

[jqnatividad]

Just upgraded polars to 0.40.0 and unfortunately, its still not returning the expected result.

[alexander-beedie]

FYI: the fix will be landing on our side in the (very) near future ;)
pola-rs/polars#16507