westonpace commented on a change in pull request #11579:
URL: https://github.com/apache/arrow/pull/11579#discussion_r745948529
##########
File path: cpp/src/arrow/compute/exec/hash_join_node.cc
##########
@@ -43,32 +43,49 @@ bool HashJoinSchema::IsTypeSupported(const DataType& type) {
return is_fixed_width(id) || is_binary_like(id) || is_large_binary_like(id);
}
-Result<std::vector<FieldRef>> HashJoinSchema::VectorDiff(const Schema& schema,
- const
std::vector<FieldRef>& a,
- const
std::vector<FieldRef>& b) {
- std::unordered_set<int> b_paths;
- for (size_t i = 0; i < b.size(); ++i) {
- ARROW_ASSIGN_OR_RAISE(auto match, b[i].FindOne(schema));
- b_paths.insert(match[0]);
+Result<std::vector<FieldRef>> HashJoinSchema::ComputePayload(
+ const Schema& schema, const std::vector<FieldRef>& output,
+ const std::vector<FieldRef>& filter, const std::vector<FieldRef>& keys) {
+ // payload = (output + filter) - keys, with no duplicates
Review comment:
Ok. I spoke to Sasha a bit about this as well. Just to close the loop,
in IR, in a join expression, field reference indices are indices into the
concatenated {left_schema, right_schema} correct? So `field_ref(5)` is
`left_schema[5]` if `len(left_schema) > 5` and otherwise it is `right_schema[5
- len(left_schema)]`?
--
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: github-unsubscr...@arrow.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
