There is a logical exception when applying time filtering with aggregate functions.

[LingweiKuang]

Dear InfluxDB Team, When I execute an aggregate function, there is a logical difference in the results depending on whether or not parentheses are added around the time field in the predicate expression. This behavior does not align with expectations.

Steps to reproduce

In the latest version of InfluxDB, if we have an organization named "organization", we can use the Influx client to execute the following commands to create a bucket and write the specified data into that bucket.

influx bucket delete --name db0
influx bucket create --name db0 --org organization

influx write \
  -b db0 \
  -o organization \
  -p ms \
  't0,t0_tag0=initTag t0_f0=1 1641024000000'

Next, we use the influx v1 shell command to enter the Influx shell backend and execute the following query statement:

# query 1
SELECT COUNT(t0_f0) FROM db0.autogen.t0 WHERE time >= 1641024000000000000 AND time <= 1641024000000000000 GROUP BY time(1800s);

# query 2
SELECT COUNT(t0_f0) FROM db0.autogen.t0 WHERE time >= 1641024000000000000 AND (time) <= 1641024000000000000 GROUP BY time(1800s);

Expected behaviour

Query 1 returned result set: 1

Query 2 returned result set: 1

Actual behaviour

Query 1 returned result set: 1

Query 2 returned result set: 1, and there are multiple zero values.

Environment info

Influx CLI dev (git: a79a2a1b825867421d320428538f76a4c90aa34c) build_date: 2024-04-16T14:34:32Z

InfluxDB OSS v2.7.11 (git: fbf5d4a)

Linux 5.15.0-94-generic x86_64

[LingweiKuang]

[davidby-influx]

If the query works correctly without parentheses, is there any reason to add them?

[LingweiKuang]

Technically, what you said is correct, but I believe this could be a potential logical issue in InfluxQL. It might be better to address this. When time is enclosed in parentheses, the query returns a large amount of invalid data, which could waste server resources. If this issue were fixed, it should improve query efficiency and system performance.