mapleFU commented on code in PR #14293:
URL: https://github.com/apache/arrow/pull/14293#discussion_r1053558528
##########
cpp/src/parquet/encoding.cc:
##########
@@ -2537,6 +2537,100 @@ class DeltaBitPackDecoder : public DecoderImpl, virtual
public TypedDecoder<DTyp
// ----------------------------------------------------------------------
// DELTA_LENGTH_BYTE_ARRAY
+// ----------------------------------------------------------------------
+// DeltaLengthByteArrayEncoder
+
+class DeltaLengthByteArrayEncoder : public EncoderImpl,
+ virtual public TypedEncoder<ByteArrayType>
{
+ public:
+ using T = typename ByteArrayType::c_type;
+
+ explicit DeltaLengthByteArrayEncoder(const ColumnDescriptor* descr,
MemoryPool* pool)
+ : EncoderImpl(descr, Encoding::DELTA_LENGTH_BYTE_ARRAY,
+ pool = ::arrow::default_memory_pool()),
+ sink_(pool),
+ length_encoder_(nullptr, pool),
+ encoded_size_{0},
+ lengths_{0} {}
+
+ std::shared_ptr<Buffer> FlushValues() override;
+
+ int64_t EstimatedDataEncodedSize() override { return encoded_size_; }
+
+ using TypedEncoder<ByteArrayType>::Put;
+
+ void Put(const ::arrow::Array& values) override;
+
+ void Put(const T* buffer, int num_values) override;
+
+ void PutSpaced(const T* src, int num_values, const uint8_t* valid_bits,
+ int64_t valid_bits_offset) override;
+
+ protected:
+ ::arrow::BufferBuilder sink_;
+ DeltaBitPackEncoder<Int32Type> length_encoder_;
+ uint32_t encoded_size_;
+ std::vector<int32_t> lengths_;
+};
+
+void DeltaLengthByteArrayEncoder::Put(const ::arrow::Array& values) {
+ auto src = values.data()->GetValues<ByteArray>(1);
+ Put(src, static_cast<int>(values.length()));
+}
+
+void DeltaLengthByteArrayEncoder::Put(const T* src, int num_values) {
+ if (num_values == 0) {
+ return;
+ }
+ lengths_.resize(num_values);
Review Comment:
> Let's stare back a bit. You don't _need_ a lengths variable at all.
Sorry for do the misleading. I just want to make clear our code style. But I
think here are two ways:
method1:
```
for every bytes:
(temp variable) length = bytes.length
Put(&length, 1)
```
or method2
```
Copy all lengths to a vector
Put(vec.data(), vec.size());
```
Seems that we prefer the
Seems, When should we buffer the vector values, like ? And when should we
just create a vector? Seems there would be a `sizeof(variable) * BatchSize` of
memory allocation.
By the way, I found there are both `ArrowPoolVector` (with arrow allocator)
and `std::vector` (with std allocator), when should we use them? @pitrou
--
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]
