diff --git a/records_mover/records/sources/table.py b/records_mover/records/sources/table.py
index a279fb78a..426e70ca2 100644
--- a/records_mover/records/sources/table.py
+++ b/records_mover/records/sources/table.py
@@ -14,7 +14,7 @@
 from ...url.resolver import UrlResolver
 from records_mover.url.base import BaseDirectoryUrl
 import logging
-from typing import Iterator, List, TYPE_CHECKING
+from typing import Generator, Iterator, List, TYPE_CHECKING
 if TYPE_CHECKING:
     from .dataframes import DataframesRecordsSource  # noqa
     from pandas import DataFrame  # noqa
@@ -89,13 +89,17 @@ def to_dataframes_source(self,
         logger.info(f"Exporting in chunks of up to {chunksize} rows by {num_columns} columns")
 
         quoted_table = quote_schema_and_table(db, self.schema_name, self.table_name)
-        chunks: Iterator['DataFrame'] = \
+        chunks: Generator['DataFrame', None, None] = \
             pandas.read_sql(f"SELECT * FROM {quoted_table}",
                             con=db,
                             chunksize=chunksize)
-        yield DataframesRecordsSource(dfs=self.with_cast_dataframe_types(records_schema, chunks),
-                                      records_schema=records_schema,
-                                      processing_instructions=processing_instructions)
+        try:
+            yield DataframesRecordsSource(dfs=self.with_cast_dataframe_types(records_schema,
+                                                                             chunks),
+                                          records_schema=records_schema,
+                                          processing_instructions=processing_instructions)
+        finally:
+            chunks.close()
 
     def with_cast_dataframe_types(self,
                                   records_schema: RecordsSchema,
diff --git a/tests/unit/records/sources/test_table.py b/tests/unit/records/sources/test_table.py
index d9b9576d6..027c6c8d8 100644
--- a/tests/unit/records/sources/test_table.py
+++ b/tests/unit/records/sources/test_table.py
@@ -35,6 +35,7 @@ def test_to_dataframes_source(self,
         mock_columns = [mock_column]
         mock_db.dialect.get_columns.return_value = mock_columns
         mock_quoted_table = mock_quote_schema_and_table.return_value
+        mock_chunks = mock_read_sql.return_value
         with self.table_records_source.to_dataframes_source(mock_processing_instructions) as\
                 df_source:
             self.assertEqual(df_source, mock_DataframesRecordsSource.return_value)
@@ -51,6 +52,8 @@ def test_to_dataframes_source(self,
                 assert_called_with(dfs=ANY,
                                    processing_instructions=mock_processing_instructions,
                                    records_schema=mock_records_schema)
+            mock_chunks.close.assert_not_called()
+        mock_chunks.close.assert_called()
 
     @patch('records_mover.records.sources.table.RecordsUnloadPlan')
     @patch('records_mover.records.sources.table.MoveResult')