Punisheroot commented on code in PR #50710:
URL: https://github.com/apache/arrow/pull/50710#discussion_r3674597158
##########
cpp/src/parquet/decoder.cc:
##########
@@ -1295,12 +1294,80 @@ class DictByteArrayDecoderImpl : public
DictDecoderImpl<ByteArrayType> {
int64_t valid_bits_offset,
typename EncodingTraits<ByteArrayType>::Accumulator*
out,
int* out_num_values) {
+ const auto* dict_values = dictionary_->data_as<ByteArray>();
+ const int values_to_decode = num_values - null_count;
+
+ switch (out->builder->type()->id()) {
+ case ::arrow::Type::BINARY:
+ case ::arrow::Type::STRING:
+ case ::arrow::Type::LARGE_BINARY:
+ case ::arrow::Type::LARGE_STRING: {
+ if (values_to_decode > 0) {
+ RETURN_NOT_OK(indices_scratch_space_->TypedResize<int32_t>(
+ values_to_decode, /*shrink_to_fit=*/false));
+ }
+ auto* decoded_indices =
indices_scratch_space_->mutable_data_as<int32_t>();
+ const int num_indices = idx_decoder_.GetBatch(decoded_indices,
values_to_decode);
+ if (ARROW_PREDICT_FALSE(num_indices != values_to_decode)) {
+ return Status::Invalid("Invalid number of indices: ", num_indices);
+ }
+
+ int64_t data_length = 0;
+ for (int i = 0; i < values_to_decode; ++i) {
+ const auto index = decoded_indices[i];
+ RETURN_NOT_OK(IndexInBounds(index));
+ if (ARROW_PREDICT_FALSE(AddWithOverflow(
+ data_length, static_cast<int64_t>(dict_values[index].len),
+ &data_length))) {
+ return Status::Invalid(
+ "excess expansion while decoding dictionary-encoded
BYTE_ARRAY");
+ }
+ }
+
+ auto append_predecoded = [&](auto* helper) {
+ int values_decoded = 0;
+ int pos_indices = 0;
+ int64_t remaining_data_length = data_length;
+
+ RETURN_NOT_OK(VisitBitRuns(
+ valid_bits, valid_bits_offset, num_values,
+ [&](int64_t position, int64_t length, bool valid) {
+ if (valid) {
+ for (int64_t i = 0; i < length; ++i) {
+ const auto& val =
dict_values[decoded_indices[pos_indices++]];
+ RETURN_NOT_OK(helper->AppendValue(
+ val.ptr, static_cast<int32_t>(val.len),
remaining_data_length));
+ remaining_data_length -= val.len;
+ }
+ values_decoded += static_cast<int>(length);
+ } else {
+ for (int64_t i = 0; i < length; ++i) {
+ helper->UnsafeAppendNull();
+ }
+ }
Review Comment:
I benchmarked this exact change by replacing the per-value loop with
`helper->AppendNulls(length)` in an ABBA comparison using the same Windows MSVC
Release build. It improved the 10% randomly distributed null case by
approximately 2.4–2.6%, but regressed the 50% null case by approximately
10–12%. The latter produces many short null runs, for which the bulk call is
not consistently beneficial.
A separate microbenchmark showed clear benefits only for sufficiently long
runs, so an adaptive run-length threshold may be worth exploring separately.
For this PR, I’d prefer to retain the existing per-value path and keep the
change focused on exact dictionary payload reservation.
--
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]
