joechenrh commented on code in PR #485:
URL: https://github.com/apache/arrow-go/pull/485#discussion_r2310543434
##########
parquet/file/column_reader.go:
##########
@@ -437,16 +446,27 @@ func (c *columnChunkReader) initDataDecoder(page Page,
lvlByteLen int64) error {
format.Encoding_DELTA_LENGTH_BYTE_ARRAY,
format.Encoding_DELTA_BINARY_PACKED,
format.Encoding_BYTE_STREAM_SPLIT:
- c.curDecoder =
c.decoderTraits.Decoder(parquet.Encoding(encoding), c.descr, false, c.mem)
- c.decoders[encoding] = c.curDecoder
+ c.curDecoder =
c.decoderTraits.Decoder(parquet.Encoding(enc), c.descr, false, c.mem)
+ c.decoders[enc] = c.curDecoder
case format.Encoding_RLE_DICTIONARY:
return errors.New("parquet: dictionary page must be
before data page")
default:
- return fmt.Errorf("parquet: unknown encoding type %s",
encoding)
+ return fmt.Errorf("parquet: unknown encoding type %s",
enc)
+ }
+ }
+
+ switch c.descr.PhysicalType() {
+ case parquet.Types.FixedLenByteArray:
+ c.curDecoder = &encoding.FixedLenByteArrayDecoderWrapper{
+ FixedLenByteArrayDecoder:
c.curDecoder.(encoding.FixedLenByteArrayDecoder),
+ }
+ case parquet.Types.ByteArray:
+ c.curDecoder = &encoding.ByteArrayDecoderWrapper{
+ ByteArrayDecoder:
c.curDecoder.(encoding.ByteArrayDecoder),
Review Comment:
> I think that's also part of why we were creating new buffers and copying
the data rather than being able to reuse the buffers in the page reader (the
thing you're changing in this PR)
But `memory.NewBufferBytes` doesn't copy the data, just ref it. That's the
reason I open this PR.
https://github.com/apache/arrow-go/blob/c6ce2ef4e55009a786cf04b3845eba5170c98066/arrow/memory/buffer.go#L50-L53
--
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]
