pitrou commented on code in PR #50164:
URL: https://github.com/apache/arrow/pull/50164#discussion_r3414439819
##########
cpp/src/arrow/compute/kernels/vector_selection_take_internal.cc:
##########
@@ -488,6 +489,123 @@ Status FixedWidthTakeExec(KernelContext* ctx, const
ExecSpan& batch, ExecResult*
namespace {
+template <typename IndexCType>
+Status VarBinaryViewTakeTyped(const ArraySpan& values, const ArraySpan&
indices,
+ BinaryViewType::c_type* out_views, uint8_t*
out_validity,
+ int64_t* valid_count) {
+ const auto* source_views = values.GetValues<BinaryViewType::c_type>(1);
+ const auto* index_values = indices.GetValues<IndexCType>(1);
+
+ const bool values_may_have_nulls = values.MayHaveNulls();
+ const bool indices_may_have_nulls = indices.MayHaveNulls();
+
+ if (!values_may_have_nulls && !indices_may_have_nulls) {
+ for (int64_t out_i = 0; out_i < indices.length; ++out_i) {
+ out_views[out_i] =
source_views[static_cast<int64_t>(index_values[out_i])];
+ }
+ *valid_count = indices.length;
+ return Status::OK();
+ }
+
+ for (int64_t out_i = 0; out_i < indices.length; ++out_i) {
+ if (indices_may_have_nulls &&
+ !bit_util::GetBit(indices.buffers[0].data, indices.offset + out_i)) {
+ continue;
+ }
+
+ const int64_t source_i = static_cast<int64_t>(index_values[out_i]);
+ const bool source_valid =
+ !values_may_have_nulls ||
+ bit_util::GetBit(values.buffers[0].data, values.offset + source_i);
+ if (!source_valid) {
+ continue;
+ }
+
+ out_views[out_i] = source_views[source_i];
+ if (out_validity != nullptr) {
+ bit_util::SetBit(out_validity, out_i);
+ }
Review Comment:
`OptionalBitmapAnd` would probably be much more efficient that settings
output validity bits individually.
--
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]
