rtpsw commented on code in PR #13880:
URL: https://github.com/apache/arrow/pull/13880#discussion_r956555941
##########
cpp/src/arrow/compute/exec/asof_join_node.cc:
##########
@@ -657,34 +986,85 @@ class AsofJoinNode : public ExecNode {
return std::make_shared<arrow::Schema>(fields);
}
+ static inline Result<col_index_t> FindColIndex(const Schema& schema,
+ const FieldRef& field_ref,
+ const util::string_view&
key_kind) {
+ auto match_res = field_ref.FindOne(schema);
+ if (!match_res.ok()) {
+ return Status::Invalid("Bad join key on table : ",
match_res.status().message());
+ }
+ auto match = match_res.ValueOrDie();
+ if (match.indices().size() != 1) {
+ return Status::Invalid("AsOfJoinNode does not support a nested ",
+ to_string(key_kind), "-key ",
field_ref.ToString());
+ }
+ return match.indices()[0];
+ }
+
static arrow::Result<ExecNode*> Make(ExecPlan* plan, std::vector<ExecNode*>
inputs,
const ExecNodeOptions& options) {
DCHECK_GE(inputs.size(), 2) << "Must have at least two inputs";
const auto& join_options = checked_cast<const
AsofJoinNodeOptions&>(options);
- ARROW_ASSIGN_OR_RAISE(std::shared_ptr<Schema> output_schema,
- MakeOutputSchema(inputs, join_options));
+ if (join_options.by_key.size() == 0) {
+ return Status::Invalid("AsOfJoin by_key must not be empty");
+ }
Review Comment:
I've looked into this and it is not a trivial change, not only because new
expected results (the empty key means all 3 joined tables will contribute a
"key" column) need to be validated but also because I ran into unclear runtime
problems on my first couple of tries. So, I'll need some time, or we could
decide to split this into a separate issue.
--
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]
