sqlglot.parse_one - use read keyword argument

[ADBond]

Type of PR

• BUG
• FEAT
• MAINT
• DOC

Is your Pull Request linked to an existing Issue or Pull Request?

Fixes #1995.

For older versions (before 17.1.0) of sqlglot, SqlglotColumnTreeBuilder.from_raw_column_name_or_column_reference
would break.

Give a brief description for the solution you have provided

in 17.1.0 the dialect keyword argument was introduced as a synonym for read. But this breaks on earlier versions of sqlglot, so this changes to use the read parameter, which is compatible with 'all' (that we support) versions of sqlglot.

As a side-point, I wonder if it makes sense for us to wrap parse_one, as we use it in several places in the codebase? That way if in the future the api changes and we need to provide compatibility code, we only need it in one place. We also avoid these kinds of issues, as we wouldn't need to add further calls (directly) to parse_one, where it would be easy to accidentally use dialect= if the developer has a newer sqlglot installed (and as used in CI).

PR Checklist

• Added documentation for changes
• Added feature to example notebooks or tutorial (if appropriate)
• Added tests (if appropriate)
• Updated CHANGELOG.md (if appropriate)
• Made changes based off the latest version of Splink
• Run the linter

[ADBond]

Test script (which breaks prior to this PR for sqlglot < 17.1.0:

expand

import pandas as pd

from splink.duckdb.linker import DuckDBLinker

df = pd.DataFrame(
    [
        {
            "unique_id": 1,
            "bothinitials": "BA",
            "sex": "M",
            "dob": "2020-01-01",
            "sector": "Q",
        },
        {
            "unique_id": 2,
            "bothinitials": "ZO",
            "sex": "M",
            "dob": "2020-03-01",
            "sector": "T",
        },
        {
            "unique_id": 3,
            "bothinitials": "BA",
            "sex": "M",
            "dob": "2020-01-01",
            "sector": "Q",
        },
        {
            "unique_id": 4,
            "bothinitials": "ZB",
            "sex": "F",
            "dob": "2020-05-01",
            "sector": "Q",
        },
        {
            "unique_id": 5,
            "bothinitials": "OK",
            "sex": "F",
            "dob": "2020-01-01",
            "sector": "Q",
        },
        {
            "unique_id": 6,
            "bothinitials": "NE",
            "sex": "F",
            "dob": "2020-06-01",
            "sector": "Q",
        },
        {
            "unique_id": 7,
            "bothinitials": "OK",
            "sex": "M",
            "dob": "2020-08-01",
            "sector": "T",
        },
        {
            "unique_id": 8,
            "bothinitials": "NE",
            "sex": "F",
            "dob": "2020-01-01",
            "sector": "Q",
        },
    ]
)

deterministic_brs = [
    # Non null match on both initials
    "l.bothinitials = r.bothinitials AND "
    # Non null match on date of birth
    "l.dob = r.dob AND "
    # Non null match on sex
    "l.sex = r.sex AND "
    # Non null match on postcode sector
    "l.sector = r.sector"
]

deterministic_settings = {
    "link_type": "dedupe_only",
    "blocking_rules_to_generate_predictions": deterministic_brs,
}
deterministic_linker = DuckDBLinker(df, deterministic_settings)