L014 triggered by leading underscore

[j-svensmark]

Search before asking

• I searched the issues and found no similar issues.

What Happened

When parsing a file with sqlfluff, I unexpectedly got a Unquoted identifiers must be consistently capitalised. error from L014. Here is an MWE

select
    _a,
    b
from foo

Expected Behaviour

I expect running sqlfluff on the above code sample shouldn't produce any error.

Observed Behaviour

When running sqlfluff lint --dialect bigquery MWE.sql I get an error from L014

== [......................../MWE.sql] FAIL
L:   3 | P:   5 | L014 | Unquoted identifiers must be consistently capitalised.
L:   4 | P:   6 | L014 | Unquoted identifiers must be consistently capitalised.
All Finished 📜 🎉!

How to reproduce

See above

Dialect

I tested with bigquery, oracle and tsql, and the issue occurred in all of them. I assume it applies to all dialects.

Version

Running in a conda env on M1 Macbook with

sqlfluff, version 1.3.0

python                    3.9.7                h88f2d9e_1

Configuration

No custom config used.

Are you willing to work on and submit a PR to address the issue?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[j-svensmark]

I suspect the logic here might be the triggering cause of this issue

sqlfluff/src/sqlfluff/rules/L010.py

Line 124 in 0feb870

if segment.raw[0] != segment.raw[0].upper():

Is it implicitly assumed that the first character is a capitalizable letter?

[alanmcruickshank]

Yep - 100% right on why this is happening. I think this should be a fairly easy fix to find the first character where casing matters and infer from that one 👍 . @j-svensmark - would you be up for submitting a PR for this?

[j-svensmark]

Sure, please see the PR linked above.
Btw, in case anyone needs a workaround for this, explicitly setting the capitalization policy in the .sqlfluff config silences this linting error

[sqlfluff:rules:L014]
extended_capitalisation_policy = lower