emkornfield commented on a change in pull request #8897:
URL: https://github.com/apache/arrow/pull/8897#discussion_r542606601
##########
File path: cpp/src/arrow/util/decimal.cc
##########
@@ -718,6 +718,48 @@ Result<Decimal256> Decimal256::FromString(const char* s) {
return FromString(util::string_view(s));
}
+Result<Decimal256> Decimal256::FromBigEndian(const uint8_t* bytes, int32_t
length) {
+ static constexpr int32_t kMinDecimalBytes = 1;
+ static constexpr int32_t kMaxDecimalBytes = 32;
+
+ std::array<uint64_t, 4> little_endian_array;
+
+ if (ARROW_PREDICT_FALSE(length < kMinDecimalBytes || length >
kMaxDecimalBytes)) {
+ return Status::Invalid("Length of byte array passed to
Decimal128::FromBigEndian ",
+ "was ", length, ", but must be between ",
kMinDecimalBytes,
+ " and ", kMaxDecimalBytes);
+ }
+
+ // Bytes are coming in big-endian, so the first byte is the MSB and
therefore holds the
+ // sign bit.
+ const bool is_negative = static_cast<int8_t>(bytes[0]) < 0;
+
+ for (int word_idx = 0; word_idx < 4; word_idx++) {
+ const int32_t word_length = std::min(length,
static_cast<int32_t>(sizeof(uint64_t)));
+
+ if (word_length == 8) {
+ // Full words can be assigned as is (and are UB with the shift below).
+ little_endian_array[word_idx] =
+ UInt64FromBigEndian(bytes + length - word_length, word_length);
+ } else {
+ // Sign extend the word its if necessary
+ uint64_t word = -1 * is_negative;
+ if (length > 0) {
+ // Incorporate the actual values if present.
+ // Shift left enough bits to make room for the incoming int64_t
+ word = SafeLeftShift(word, word_length * CHAR_BIT);
+ // Preserve the upper bits by inplace OR-ing the int64_t
+ word |= UInt64FromBigEndian(bytes + length - word_length, word_length);
+ }
+ little_endian_array[word_idx] = word;
+ }
+ // Move on to the next word.
+ length -= sizeof(uint64_t);
Review comment:
yes, I added a comment. do you think this is a bug? Is there a way of
framing this logic that is less surprising to you?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
