SNOW-1269820: Structured Type support for pandas and iceberg. (#1907)

Author: sfc-gh-jrose

DESCRIPTION.md

@@ -16,6 +16,7 @@ Source code is also available at: https://github.com/snowflakedb/snowflake-conne
 
   - Added easy logging configuration so that users can easily generate log file by setup log config in `$SNOWFLAKE_HOME/config.toml`.
   - Improved s3 acceleration logic when connecting to China endpoint.
+  - Added support for structured types to fetch_pandas_all
 
 - v3.8.1(April 09, 2024)
 

src/snowflake/connector/cursor.py

@@ -1314,7 +1314,17 @@ def fetch_pandas_batches(self, **kwargs: Any) -> Iterator[DataFrame]:
         return self._result_set._fetch_pandas_batches(**kwargs)
 
     def fetch_pandas_all(self, **kwargs: Any) -> DataFrame:
-        """Fetch Pandas dataframes in batches, where 'batch' refers to Snowflake Chunk."""
+        """
+        Fetch Pandas dataframes in batches, where 'batch' refers to Snowflake Chunk.
+
+        Returns:
+            A pandas dataframe containing the cursors query results.
+
+        Note:
+            Timestamp types that are nested in structured types may not be accurately represented
+            due to a limitation in the pyarrow to pandas conversion.
+            See here for more information: https://github.com/apache/arrow/issues/4116
+        """
         self.check_can_use_pandas()
         if self._prefetch_hook is not None:
             self._prefetch_hook()