events.duration() fails after 1000 rows

[Ogoku]

Hi all and @tomhollingworth,

I have a strange behaviour using events.duration( )

Every 1000th row events.duration(unit: 1m, stopColumn: "_stop" ) is not calculating current row with the next one. Instead it is calculating the differences to the given stop time.

(1222min ~20h which is _time to _stop)

When I export my query via csv I can see that every 1000th entry the result ist _time to _stop.

Steps to reproduce:
Whole query:

`
import "contrib/tomhollingworth/events"

from(bucket: "statistik_daten")
|> range(start: v.timeRangeStart, stop: v.timeRangeStop)
|> filter(fn: (r) => r["_measurement"] == "opc_ua_mes_2")
|> filter(fn: (r) => r["maschinen_name"] == "${Maschinen_Name}")
|> events.duration(unit: 1ms, stopColumn: "_stop" )
`

Read more than 1000 entries

Expected behavior:
Except the last row, events.Duration() should always calculates the difference between current record with the next one

Actual behavior:
After 1000 rows events.Duration() calculates difference between current record and _stop value

Environment info:
Running on a Windows OS

• InfluxDB version: InfluxDB 2.0.7 (git: 2a45f0c) build_date: 2021-06-04T19:17:40Z

[williamhbaker]

This should be fixed by influxdata/flux#4064 with flux v0.139.0. This version of flux will go out with our next minor release, influxdb 2.1, which should be coming out in a few days. Would you be able to upgrade to 2.1 when it's available? If so it would be nice to see if the bug you are seeing is indeed fixed with the latest version of flux.

[Ogoku]

Hey @wbaker85. Updated Influxdb to 2.1 this morning. Problem solved!