GoGoWen commented on a change in pull request #6965:
URL: https://github.com/apache/incubator-doris/pull/6965#discussion_r739820481
##########
File path: be/src/olap/reader.cpp
##########
@@ -130,10 +130,16 @@ OLAPStatus Reader::init(const ReaderParams& read_params) {
<< ", version:" << read_params.version;
return res;
}
-
+ // optimize for single rowset reading without do aggregation when reading
all columns,
+ // and otherwise should use _agg_key_next_row for AGG_KEYS
if (_optimize_for_single_rowset(rs_readers)) {
- _next_row_func = _tablet->keys_type() == AGG_KEYS ?
&Reader::_direct_agg_key_next_row
- :
&Reader::_direct_next_row;
+ if(_tablet->keys_type() == AGG_KEYS && _return_columns.size() ==
_tablet->tablet_schema().num_columns()) {
+ _next_row_func = &Reader::_direct_agg_key_next_row;
+ } else if (_tablet->keys_type() == AGG_KEYS) {
+ _next_row_func = &Reader::_agg_key_next_row;
Review comment:
For example, the table test (k int ,k1 int ,k2 int ,k3 int, v int sum),
if we read all columns (k,k1,k2,k3,v), then it definitely no need to call
agg. However, if we just need (k,k1,v), then if we do not call agg, there
are many many duplicate-key rows for next caller(may be aggregation node),
it's not a good idea to pass these rows without aggregation to aggregation node
for a query.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
