YAML value column selector

[mark-druffel]

Description

I can't get pb.col() to work within the YAML workflow. It appears, for col_vals_ge() & col_vals_gt() at least, pb.col() inputs fail in _string_date_dttm_conversion(value=value).

Reproducible example

Create some data for example:

import pointblank as pb
import polars as pl
from datetime import date

df = pl.DataFrame(
    {
        "id": list(range(123, 123 + 6)),  # 123, 124, 125, 126, 127, 128
        "start_date": [
            date(2024, 5, 1),
            date(2024, 6, 1),
            date(2024, 5, 15),
            date(2024, 7, 1),
            date(2024, 6, 10),
            date(2024, 5, 20),
        ],
        "end_date": [
            date(2024, 5, 10),
            date(2024, 6, 5),
            date(2024, 5, 25),
            date(2024, 7, 10),
            date(2024, 6, 5),
            date(2024, 5, 30),
        ],
    }
)

This validation works, 1 row fails because the end date is later than the start date:

validation = (
    pb.Validate(data=df)
    .col_vals_ge(columns="end_date", value=pb.col("start_date"))
    .interrogate()
)
validation

This validation fails, I get the error ValueError: If value= is provided as a string it must be a date or datetime string:

# Set to small_table because None fails due to validator that runs before set_tbl, allowing None would be helpful but totally separate issue :) 
pb_yaml = '''
tbl: small_table
steps:
- col_vals_ge:
    columns: [end_date]
    value: pb.col("start_date")
'''
result = pb.yaml_interrogate(pb_yaml, set_tbl = df)

Expected result

I want the YAML validation to provide the same output as the python code version above.

Development environment

macOS 14.5
pointblank 0.14.0

[rich-iannone]

Thanks for reporting (and extra thanks for the example!). I'm come up with a fix for this shortly (just getting back to Pointblank dev work more recently).