Method FillForLast

cpp/src/parquet/arrow/path_internal.cc:386–433

 private:
  IterationResult FillForLast(ElementRange* range, ElementRange* child_range,
                              PathWriteContext* context) {
    // First fill int the remainder of the list.
    RETURN_IF_ERROR(FillRepLevels(child_range->Size(), rep_level_, context));

    // Once we've reached this point the following preconditions should hold:
    // 1.  There are no more repeated path nodes to deal with.
    // 2.  Null values would have shortened the range to ensure all remaining
    //     list elements are present (though they may be empty lists).
    // 3.  No element of range spans a parent list (intermediate
    //     list nodes only handle one list entry at a time).
    //
    // Given these preconditions it is safe to fill runs on contiguous non-empty
    // lists here and expand the range in the child node accordingly.
    while (!range->Empty()) {
      ElementRange next_child_range = selector_.GetRange(range->start);
      if (next_child_range.Empty()) {
        // The empty range will need to be handled after we pass down the accumulated
        // range because it affects def_level placement and we need to get the children
        // def_levels entered first.
        break;
      }
      // FillForLast extends child_range by updating only its end. Non-contiguous
      // selectors must split at gaps.
      if constexpr (RangeSelector::kContiguous) {
        DCHECK_EQ(next_child_range.start, child_range->end)
            << next_child_range.start << " != " << child_range->end;
      } else {
        if (next_child_range.start != child_range->end) {
          break;
        }
      }
      // This is the start of a new list. We can be sure it only applies
      // to the previous list (and doesn't jump to the start of any list
      // further up in nesting due to the constraints mentioned at the start
      // of the function).
      RETURN_IF_ERROR(context->AppendRepLevel(prev_rep_level_));
      RETURN_IF_ERROR(context->AppendRepLevels(next_child_range.Size() - 1, rep_level_));
      child_range->end = next_child_range.end;
      ++range->start;
    }

    // Do book-keeping to track the elements of the arrays that are actually visited
    // beyond this point.  This is necessary to identify "gaps" in values that should
    // not be processed (written out to parquet).
    context->RecordPostListVisit(*child_range);
    return kNext;
  }

  RangeSelector selector_;
  int16_t prev_rep_level_;