JakeDern commented on code in PR #10128:
URL: https://github.com/apache/arrow-rs/pull/10128#discussion_r3839986591
##########
arrow-ipc/src/writer.rs:
##########
@@ -949,29 +960,48 @@ impl IpcDataGenerator {
ipc_write_context: &mut IpcWriteContext,
sink: &mut S,
) -> Result<IpcWriteMetadata, ArrowError> {
- let encoded_dictionaries =
- self.encode_all_dicts(batch, dictionary_tracker, write_options,
ipc_write_context)?;
-
- let mut dictionary_block_sizes =
Vec::with_capacity(encoded_dictionaries.len());
- for dict in encoded_dictionaries {
- dictionary_block_sizes.push(sink.write_encoded_data(dict,
write_options)?);
- }
+ let dictionaries = self.collect_all_dicts(batch, dictionary_tracker,
write_options)?;
+ let mut dictionary_block_sizes =
Vec::with_capacity(dictionaries.len());
let capacity = batch
.columns()
.iter()
.map(|a| estimate_encoded_buffer_count(a.data_type()))
.sum();
let mut encoded_buffers: Vec<EncodedBuffer> =
Vec::with_capacity(capacity);
- let (metadata, body_len, tail_pad) = self.record_batch_to_bytes(
+
+ for dict in &dictionaries {
+ encoded_buffers.clear();
+
+ let body_len = self.dictionary_batch_to_sink(
+ dict,
+ write_options,
+ ipc_write_context,
+ &mut IpcBodySink::Collect(&mut encoded_buffers),
+ )?;
+
+ dictionary_block_sizes.push(sink.write_record_batch(
+ ipc_write_context.mut_fbb().finished_data(),
+ &mut encoded_buffers,
+ body_len,
+ write_options,
+ )?);
+ }
+
+ encoded_buffers.clear();
+ let body_len = self.record_batch_to_bytes(
batch,
write_options,
ipc_write_context,
&mut IpcBodySink::Collect(&mut encoded_buffers),
)?;
- let (padded_header_len, body_len) =
- sink.write_record_batch(metadata, encoded_buffers, body_len,
tail_pad, write_options)?;
+ let (padded_header_len, body_len) = sink.write_record_batch(
+ ipc_write_context.mut_fbb().finished_data(),
+ &mut encoded_buffers,
+ body_len,
+ write_options,
+ )?;
Review Comment:
Well, I'm at least glad I'm not losing my mind and it's not obvious to you
either 😅
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
