feat(typing): add databricks type inference for REGEXP_INSTR, REGEXP_LIKE, REGEXP_SUBSTR, REGR_AVGX

[fivetran-amrutabhimsenayachit]

Summary

Adds Databricks type inference support for REGEXP_INSTR (INT), REGEXP_LIKE (BOOLEAN via fixture), REGEXP_SUBSTR (VARCHAR via parser mapping to RegexpExtract), and REGR_AVGX (DOUBLE), plus fixture coverage for all four functions.

Tickets

• RD-1229633 (REGEXP_INSTR) — new INT entry in typing/databricks.py
• RD-1229634 (REGEXP_LIKE) — fixture entry only (inherited from base typing)
• RD-1229635 (REGEXP_REPLACE) — already complete, no changes
• RD-1229636 (REGEXP_SUBSTR) — parser mapping REGEXP_SUBSTR→RegexpExtract + fixture entry
• RD-1229637 (REGR_AVGX) — new DOUBLE entry in typing/databricks.py

Test plan

• make style — PASS
• make unit — PASS

[github-actions]

SQLGlot Integration Test Results

✅ All tests passed

Comparing:

• this branch (sqlglot:type-inference-batch-2 @ sqlglot 2195638)
• baseline (main @ sqlglot fd6d4d6)

By Dialect

dialect	main	feature branch	transitions	links
databricks -> databricks	9982/11820 passed (84.5%)	9998/11820 passed (84.6%)	16 fail -> pass	full result / delta

Overall

main: 192428 total, 153523 passed (pass rate: 79.8%)

sqlglot:type-inference-batch-2: 180234 total, 142394 passed (pass rate: 79.0%)

Transitions:
16 fail -> pass

Dialect pair changes: 0 previous results not found, 3 current results not found

✅ All tests passed

[geooo109]

Let's also add tests containing the ALL, DISTINCT, andREGR_AVGX(...) OVER (PARTITION BY 1)

[geooo109]

Can you check if we can reuse exp.RegexpExtract for this? The default group should be 0 here. So, check the semantics for regexp_extract vs regexp_substr, if they match (for group 0), we can reuse the existing expression. If they don't match, let's keep this addition and add round-trip tests.

Minimal example:
SELECT regexp_extract('Order: 100-200', '(\\d+)-(\\d+)', 0)
> 100-200

SELECT regexp_substr('Order: 100-200', '(\\d+)-(\\d+)')
> 100 -200