To support schema evolution, Iceberg add schema information to Parquet file metadata. But for early iceberg version, it doesn't write any schema information to Parquet file. This PR is to support read parquet without schema information.