-
Notifications
You must be signed in to change notification settings - Fork 3.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
RC30 WHERE time filter Regression #2557
Comments
Thanks @otoolep can you clarify how I should change the specific query I listed above to be the proper syntax? select sum(value) from bps_out WHERE time > now() - 1m AND pop =~ /new_york/ GROUP BY time(60s) fill |
Actually @Jhors2 -- I see you have a @corylanou -- should the query as is work? It has the |
@otoolep yes this should work. Let me see why it is failing this use case. Looks like some more tests are in order. |
I'm unable to replicate this. Using a query exactly like they above (accept to use my current data) this works on RC30 for me. > select sum(value) from mem WHERE time > now() - 1m AND host =~ /server01/ GROUP BY time(60s) fill(0)
name: mem
---------
time sum
2015-05-13T14:38:00Z 0
2015-05-13T14:39:00Z 0 |
After more testing on my side it appears the parser works correctly for some metrics but not all. Specifically the metrics that don't appear to be working properly I am writing with a microsecond epoch timestamp and identifying the precision as such. If there is a way I can provide my measurements and exact data to the Influx team (preferrably offline) I would love to find root cause. |
The precision is currently declared per point, not batched. The only batching I am doing is Database declaration and Retention policy. |
@Jhors2 if acceptable you can email your sample data to support@influxdb.com. |
I ran into this too and we had to downgrade back to rc29. I found that using the However when I updated our Grafana graphs to change the ordering, I found that in practice most of them just hung forever until they timed out. The "bad" queries would return a 400 immediately, but the rewritten queries never even showed up in the InfluxDB log as completing. |
Can you send me specific queries that do not work? The data will have nothing to do with the parser, as this fires before we even send to the query engine. |
I have sent some sample data that I am seeing the issue with to the influx support email. Continues to recreate that problem for me, also the below query recreates the issue: |
I am seeing the same issue, re-ordering the where clause to put the time at the end seems to work for me.
but this does not:
FWIW, I found that if I remove one of the 'and' clauses it works for whatever reason (although not helpful cause I need all the ands):
|
@dneuman64 thx for the info. I'm able to replicate this now. Looking into it now. |
Hello,
I upgraded from RC29 to RC30 this morning. Since doing so I am hitting the following parser error for a query that worked in RC29, the query is as follows:
select sum(value) from bps_out WHERE time > now() - 1m AND pop =~ /new_york/ GROUP BY time(60s) fill(0)
This is returning the following error:
error: "error parsing query: aggregate functions with GROUP BY time require a WHERE time clause"
I already changed all of my writers to use the new "time" nomenclature away from "timestamp".
The text was updated successfully, but these errors were encountered: