feat(databricks,duckdb,postgres,spark,snowflake)!: ARRAY_APPEND null propagation

[fivetran-MichaelLee]

The ARRAY_APPEND function handles NULL input arrays differently across the dialects that support it:

Database	Behavior	Result
Snowflake	Returns NULL	NULL
DuckDB	Creates new array	[element]
PostgreSQL	Creates new array	{element}
Databricks	Returns NULL	NULL
Spark	Returns NULL	NULL

This PR introduces a new arg_types parameter null_propagation to preserve NULL handling semantics when transpiling between these dialects.

[github-actions]

SQLGlot Integration Test Results

Comparing:

• this branch (sqlglot:mlee/transpile-array_append, sqlglot version: mlee/transpile-array_append)
• baseline (main, sqlglot version: 28.6.1.dev44)

⚠️ Limited to dialects: bigquery, duckdb, snowflake

By Dialect

dialect	main	sqlglot:mlee/transpile-array_append	difference	links
bigquery -> bigquery	2592/2624 passed (98.8%)	2592/2624 passed (98.8%)	No change	full result / delta
bigquery -> duckdb	1846/2623 passed (70.4%)	1845/2623 passed (70.3%)	No change	full result / delta
duckdb -> duckdb	4003/4003 passed (100.0%)	4003/4003 passed (100.0%)	No change	full result / delta
snowflake -> duckdb	615/847 passed (72.6%)	615/847 passed (72.6%)	No change	full result / delta
snowflake -> snowflake	847/847 passed (100.0%)	847/847 passed (100.0%)	No change	full result / delta

Overall

main: 10944 total, 9903 passed (pass rate: 90.5%), sqlglot version: 28.6.1.dev44

sqlglot:mlee/transpile-array_append: 10944 total, 9902 passed (pass rate: 90.5%), sqlglot version: mlee/transpile-array_append

Difference: No change

[VaggelisD]

Nice work @fivetran-MichaelLee!

[georgesittas]

Nice work @fivetran-MichaelLee!