lidavidm commented on a change in pull request #11853:
URL: https://github.com/apache/arrow/pull/11853#discussion_r777548551
##########
File path: cpp/src/arrow/compute/kernels/vector_replace.cc
##########
@@ -442,23 +442,414 @@ struct ReplaceWithMaskFunctor {
}
return ReplaceWithMask<Type>::ExecArrayMask(ctx, array, mask,
replacements, output);
}
+
+ static std::shared_ptr<KernelSignature> GetSignature(detail::GetTypeId
get_id) {
+ return KernelSignature::Make(
+ {InputType::Array(get_id.id), InputType(boolean()),
InputType(get_id.id)},
+ OutputType(FirstType));
+ }
};
-} // namespace
+// This is for fixed-size types only
+template <typename Type>
+void FillNullInDirectionImpl(const ArrayData& current_chunk, const uint8_t*
null_bitmap,
+ ArrayData* output, int8_t direction,
+ const ArrayData& last_valid_value_chunk,
+ int64_t* last_valid_value_offset) {
+ uint8_t* out_bitmap = output->buffers[0]->mutable_data();
+ uint8_t* out_values = output->buffers[1]->mutable_data();
+ arrow::internal::CopyBitmap(current_chunk.buffers[0]->data(),
current_chunk.offset,
+ current_chunk.length, out_bitmap,
output->offset);
+ ReplaceWithMask<Type>::CopyData(*current_chunk.type, out_values,
+ /*out_offset=*/output->offset, current_chunk,
+ /*in_offset=*/0, current_chunk.length);
+
+ bool has_fill_value = *last_valid_value_offset != -1;
+ int64_t write_offset = direction == 1 ? 0 : current_chunk.length - 1;
+ int64_t bitmap_offset = 0;
+
+ arrow::internal::OptionalBitBlockCounter counter(null_bitmap, output->offset,
+ current_chunk.length);
+ bool use_current_chunk = false;
+ while (bitmap_offset < current_chunk.length) {
+ BitBlockCount block = counter.NextBlock();
+ if (block.AllSet()) {
+ *last_valid_value_offset =
+ write_offset + direction * (block.length - 1 + bitmap_offset);
+ has_fill_value = true;
+ use_current_chunk = true;
+ } else {
+ uint64_t block_start_offset = write_offset + direction * bitmap_offset;
+ uint64_t write_value_offset = block_start_offset;
+ if (block.popcount) {
+ for (int64_t i = 0; i < block.length; i++, write_value_offset +=
direction) {
+ auto current_bit = bit_util::GetBit(null_bitmap, bitmap_offset + i);
+ if (!current_bit) {
+ if (has_fill_value) {
+ ReplaceWithMask<Type>::CopyData(
+ *current_chunk.type, out_values, write_value_offset,
+ use_current_chunk ? current_chunk : last_valid_value_chunk,
+ *last_valid_value_offset,
+ /*length=*/1);
+ bit_util::SetBitTo(out_bitmap, write_value_offset, true);
+ }
+ } else {
+ has_fill_value = true;
+ use_current_chunk = true;
+ *last_valid_value_offset = write_value_offset;
+ }
+ }
+ } else {
+ for (int64_t i = 0; i < block.length; i++, write_value_offset +=
direction) {
+ if (has_fill_value) {
+ ReplaceWithMask<Type>::CopyData(
+ *current_chunk.type, out_values, write_value_offset,
+ use_current_chunk ? current_chunk : last_valid_value_chunk,
+ *last_valid_value_offset,
+ /*length=*/1);
+ bit_util::SetBitTo(out_bitmap, write_value_offset, true);
+ }
+ }
+ }
+ }
+ bitmap_offset += block.length;
+ }
+ output->null_count = -1;
+ output->GetNullCount();
+}
-const FunctionDoc replace_with_mask_doc(
- "Replace items selected with a mask",
- ("Given an array and a boolean mask (either scalar or of equal length),\n"
- "along with replacement values (either scalar or array),\n"
- "each element of the array for which the corresponding mask element is\n"
- "true will be replaced by the next value from the replacements,\n"
- "or with null if the mask is null.\n"
- "Hence, for replacement arrays, len(replacements) == sum(mask == true)."),
- {"values", "mask", "replacements"});
+template <typename Type, typename Enable = void>
+struct FillNullExecutor {};
-void RegisterVectorReplace(FunctionRegistry* registry) {
- auto func = std::make_shared<VectorFunction>("replace_with_mask",
Arity::Ternary(),
- &replace_with_mask_doc);
+template <typename Type>
+struct FillNullExecutor<Type, enable_if_boolean<Type>> {
+ static Status ExecFillNull(KernelContext* ctx, const ArrayData& array,
+ const uint8_t* reversed_bitmap, ArrayData* output,
+ int8_t direction, const ArrayData&
last_valid_value_chunk,
+ int64_t* last_valid_value_offset) {
+ FillNullInDirectionImpl<Type>(array, reversed_bitmap, output, direction,
+ last_valid_value_chunk,
last_valid_value_offset);
+ return Status::OK();
+ }
+};
+
+template <typename Type>
+struct FillNullExecutor<
+ Type, enable_if_t<is_number_type<Type>::value ||
+ std::is_same<Type, MonthDayNanoIntervalType>::value>> {
+ static Status ExecFillNull(KernelContext* ctx, const ArrayData& array,
+ const uint8_t* reversed_bitmap, ArrayData* output,
+ int8_t direction, const ArrayData&
last_valid_value_chunk,
+ int64_t* last_valid_value_offset) {
+ FillNullInDirectionImpl<Type>(array, reversed_bitmap, output, direction,
+ last_valid_value_chunk,
last_valid_value_offset);
+ return Status::OK();
+ }
+};
+
+template <typename Type>
+struct FillNullExecutor<Type, enable_if_fixed_size_binary<Type>> {
+ static Status ExecFillNull(KernelContext* ctx, const ArrayData& array,
+ const uint8_t* reversed_bitmap, ArrayData* output,
+ int8_t direction, const ArrayData&
last_valid_value_chunk,
+ int64_t* last_valid_value_offset) {
+ FillNullInDirectionImpl<Type>(array, reversed_bitmap, output, direction,
+ last_valid_value_chunk,
last_valid_value_offset);
+ return Status::OK();
+ }
+};
+
+template <typename Type>
+struct FillNullExecutor<Type, enable_if_base_binary<Type>> {
+ using offset_type = typename Type::offset_type;
+ using BuilderType = typename TypeTraits<Type>::BuilderType;
+
+ static Status ExecFillNull(KernelContext* ctx, const ArrayData&
current_chunk,
+ const uint8_t* reversed_bitmap, ArrayData* output,
+ int8_t direction, const ArrayData&
last_valid_value_chunk,
+ int64_t* last_valid_value_offset) {
+ BuilderType builder(current_chunk.type, ctx->memory_pool());
+ RETURN_NOT_OK(builder.Reserve(current_chunk.length));
+ RETURN_NOT_OK(builder.ReserveData(current_chunk.buffers[2]->size()));
+ int64_t array_value_index = direction == 1 ? 0 : current_chunk.length - 1;
+ const uint8_t* data = current_chunk.buffers[2]->data();
+ const uint8_t* data_prev = last_valid_value_chunk.buffers[2]->data();
+ const offset_type* offsets = current_chunk.GetValues<offset_type>(1);
+ const offset_type* offsets_prev =
last_valid_value_chunk.GetValues<offset_type>(1);
+
+ bool has_fill_value_last_chunk = *last_valid_value_offset != -1;
+ bool has_fill_value_current_chunk = false;
+ /*tuple for store: <use current_chunk(true) or last_valid_chunk(false),
+ * start offset of the current value, end offset for the current value>*/
+ std::vector<std::tuple<bool, int64_t, int64_t>> offsets_reversed;
+ RETURN_NOT_OK(VisitNullBitmapInline<>(
+ reversed_bitmap, output->offset, current_chunk.length,
+ current_chunk.GetNullCount(),
+ [&]() {
+ const offset_type offset0 = offsets[array_value_index];
+ const offset_type offset1 = offsets[array_value_index + 1];
+ offsets_reversed.push_back(
+ std::make_tuple(/*current_chunk=*/true, offset0, offset1 -
offset0));
+ *last_valid_value_offset = array_value_index;
+ has_fill_value_current_chunk = true;
+ has_fill_value_last_chunk = false;
+ array_value_index += direction;
+ return Status::OK();
+ },
+ [&]() {
+ if (has_fill_value_current_chunk || has_fill_value_last_chunk) {
Review comment:
Ah sorry, I see the comments now.
For here: instead of
```
if (A || B) {
if (!B) // A
else // B
} else {
// C
}
```
why not
```
if (A) // A
else if (B) // B
else // C
```
?
--
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]
