RalfJL commented on issue #12949:
URL: https://github.com/apache/pinot/issues/12949#issuecomment-2100456131
Different result set on successive calls of the same SQL. This might be the
same root cause as above.
So I have a complicated call
```
select b.ts, a.ts, a.callingstationid, b.callingstationid, cast(a.ts as
bigint)-cast(b.ts as bigint) as Zeit, a.ts as Stop, b.ts as Start,
a.acctuniquesessionid, b.acctuniquesessionid from radius_stop a
inner join radius_start b
on a.acctuniquesessionid = b.acctuniquesessionid and a.callingstationid =
b.callingstationid and b.ts < a.ts
where a._3gppimsi = 'null' and
not exists ( select 1 from radius_start c where c.acctuniquesessionid =
a.acctuniquesessionid and cast(c.ts as BIGINT) between cast(b.ts as BIGINT) +1
and cast(a.ts as BIGINT) -1) and
not exists ( select 1 from radius_stop c where c.acctuniquesessionid =
b.acctuniquesessionid and cast(c.ts as BIGINT) between cast(b.ts as BIGINT) +1
and cast(a.ts as BIGINT) -1) and
a.callingstationid = '94-58-CB-C1-5E-D1'
```
and the number of rows returned differ between 10 and 24. So sometimes it
returns 10 immediatley after that 17 rows, immediatley after it 24 rows than
mabye 10 again or maybe 14, ...
But there are only a distinct number of rows returned: 10, 14, 17, 21, 24
and they are not equally distributed. 10 is more often, 21 or 24 is seldom.
With every call to the same SQL the number of rows returned changes or not.
I have only 1 broker, 1 server, 1 controller
The stageStats are the same except the number of "emittedRows" and the time.
e.g.
```
"children": [
{
"type": "MAILBOX_SEND",
"executionTimeMs": 688,
"emittedRows": 17,
```
compared to
```
"children": [
{
"type": "MAILBOX_SEND",
"executionTimeMs": 632,
"emittedRows": 10,
```
The database is currently static. No rows added or deleted.
My first observation might be false. It might not have to do with the where
clause but with the fact that successive calls return different result sets
--
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]
