feat(duckdb): Fix and updates tests for pivot subquery string lietral column names

[fivetran-amrutabhimsenayachit]

Problem

Transpiling a Snowflake PIVOT query with string literal IN-values (e.g. IN ('JAN', 'FEB')) on a subquery source failed in DuckDB with zero-length delimited identifier after the optimizer ran.

Two compounding root causes:

The Snowflake parser stored pivot column names with surrounding apostrophes ('JAN' stored as identifier "'JAN'"), because it used fld.sql() instead of fld.alias_or_name. This propagated through qualify_columns star-expansion into column references like ""."'JAN'" — the 'JAN' part was invalid, and the "" empty table qualifier (cause #2) was also rejected.

When the pivot sat on a subquery with no alias, pivot.alias was '', so exp.column(name, table='') generated a zero-length delimited identifier "" as the table qualifier, which DuckDB rejects outright.

[github-actions]

SQLGlot Integration Test Results

✅ All tests passed

Comparing:

• this branch (sqlglot:RD-1205936-transpile-bug-pivot-subquery-string-literal-column-names @ sqlglot a199021)
• baseline (main @ sqlglot 6f99d4e)

Overall

main: 192441 total, 153536 passed (pass rate: 79.8%)

sqlglot:RD-1205936-transpile-bug-pivot-subquery-string-literal-column-names: 180247 total, 142391 passed (pass rate: 79.0%)

Transitions:
No change

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

✅ All tests passed

[georgesittas]

@fivetran-amrutabhimsenayachit did you check the test failures?