Skip to content

Commit 98ac3a1

Browse files
committed
Refine zero-copy streaming documentation by removing outdated information about eager conversion, emphasizing on-demand batch processing to prevent memory exhaustion.
1 parent ccc8633 commit 98ac3a1

File tree

1 file changed

+3
-5
lines changed

1 file changed

+3
-5
lines changed

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

Lines changed: 3 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -153,11 +153,9 @@ Zero-copy streaming to Arrow-based Python libraries
153153
---------------------------------------------------
154154

155155
DataFusion DataFrames implement the ``__arrow_c_stream__`` protocol, enabling
156-
zero-copy, lazy streaming into Arrow-based Python libraries. Earlier versions
157-
eagerly converted the entire DataFrame when exporting to Python Arrow APIs,
158-
which could exhaust memory on large results. With the streaming protocol,
159-
batches are produced on demand so you can process arbitrarily large results
160-
without out-of-memory errors.
156+
zero-copy, lazy streaming into Arrow-based Python libraries. With the streaming
157+
protocol, batches are produced on demand so you can process arbitrarily large
158+
results without out-of-memory errors.
161159

162160
.. note::
163161

0 commit comments

Comments
 (0)