docs: clarify SessionContext retention in DataFrame and stream consumption

Author: kosiew

docs/source/user-guide/dataframe/index.rst

@@ -168,6 +168,9 @@ out-of-memory errors.
     for batch in reader:
         ...  # process each batch as it is produced
 
+Note that streams retain the originating ``SessionContext`` internally, so the
+context can be safely dropped once the stream has been obtained.
+
 DataFrames are also iterable, yielding :class:`datafusion.RecordBatch` objects
 that implement the Arrow C data interface. These batches can be consumed by
 libraries like PyArrow without copying:

python/datafusion/dataframe.py

@@ -1116,6 +1116,11 @@ def __arrow_c_stream__(self, requested_schema: object | None = None) -> object:
         provided, only straightforward projections such as column selection or
         reordering are applied.
 
+        The returned capsule holds a reference to the originating
+        :class:`SessionContext`, keeping it alive until the stream is fully
+        consumed. This makes it safe to drop the original context after obtaining
+        the stream.
+
         Args:
             requested_schema: Attempt to provide the DataFrame using this schema.