Function is_parquet_schema_match_source_schema 

pub fn is_parquet_schema_match_source_schema(
    arrow_data_type: &DataType,
    rw_data_type: &DataType,
) -> bool

This function checks whether the schema of a Parquet file matches the user-defined schema in RisingWave. It handles the following special cases:

• Arrow’s timestamp(_, None) types (all four time units) match with RisingWave’s Timestamp type.
• Arrow’s timestamp(_, Some) matches with RisingWave’s Timestamptz type.
• Since RisingWave does not have an UInt type:
  • Arrow’s UInt8 matches with RisingWave’s Int16.
  • Arrow’s UInt16 matches with RisingWave’s Int32.
  • Arrow’s UInt32 matches with RisingWave’s Int64.
  • Arrow’s UInt64 matches with RisingWave’s Decimal.
• Arrow’s Float16 matches with RisingWave’s Float32.

Nested data type matching:

• Struct: Arrow’s Struct type matches with RisingWave’s Struct type recursively, requiring that all expected fields exist and match by name and type. Extra Arrow fields are allowed, but a declared name matching multiple Arrow siblings is ambiguous and rejected.
• List: Arrow’s List type matches with RisingWave’s List type recursively, requiring the same element type.
• Map: Arrow’s Map type matches with RisingWave’s Map type recursively, requiring the key and value types to match, and the inner struct must have exactly two fields named “key” and “value”.

Nested positions recurse through is_parquet_field_match_source_schema so a variant extension is recognized at any depth.