Function CheckIndexBoundsImpl

cpp/src/arrow/util/int_util.cc:530–561

template <typename IndexCType, bool IsSigned = std::is_signed<IndexCType>::value>
static Status CheckIndexBoundsImpl(const ArraySpan& values, uint64_t upper_limit) {
  // For unsigned integers, if the values array is larger than the maximum
  // index value (e.g. especially for UINT8 / UINT16), then there is no need to
  // boundscheck.
  if (!IsSigned &&
      upper_limit > static_cast<uint64_t>(std::numeric_limits<IndexCType>::max())) {
    return Status::OK();
  }

  const IndexCType* values_data = values.GetValues<IndexCType>(1);
  const uint8_t* bitmap = values.buffers[0].data;
  auto IsOutOfBounds = [&](IndexCType val) -> bool {
    return ((IsSigned && val < 0) ||
            (val >= 0 && static_cast<uint64_t>(val) >= upper_limit));
  };
  return VisitSetBitRuns(
      bitmap, values.offset, values.length, [&](int64_t offset, int64_t length) {
        bool block_out_of_bounds = false;
        for (int64_t i = 0; i < length; ++i) {
          block_out_of_bounds |= IsOutOfBounds(values_data[offset + i]);
        }
        if (ARROW_PREDICT_FALSE(block_out_of_bounds)) {
          for (int64_t i = 0; i < length; ++i) {
            if (IsOutOfBounds(values_data[offset + i])) {
              return Status::IndexError("Index ", ToChars(values_data[offset + i]),
                                        " out of bounds");
            }
          }
        }
        return Status::OK();
      });
}

/// \brief Branchless boundschecking of the values. Processes batches of
/// values at a time and shortcircuits when encountering an out-of-bounds