Hi, when using history sets in a connector, when will gaps be introduced in the trending?
I'm working on a connector that is reading report files and storing the data with the timestamps from the files using the "historyset" functionality. When I look at the resulting trend graph, there are a number of gaps in the trend for which I do not know what causes them. I have tried both with and without the "NT_CLOSE_HISTORY_TREND_WINDOW" call in my connector, but this does not appear to make any difference. My test data is about 2 weeks old, so only averaged data will be shown. The frequency of measurements is also quite irregular, ranging from multiple times per hour to several days apart. I was expecting a continuous trend line until the next measurement point, but that does not seem to be the case. 
Michiel,
Hard to tell where the issue may be with history sets, but below are some possibilities that we have seen impacting history sets:
- Could it be that some data points coincide with the exception values defined? https://community.dataminer.services/documentation/average-trending-calculation
- Could some timestamps be mixed up where newer values have older timestamp? This would make an unexpected set in the past where there was already data stored in the DB
- Are sets performed even when there is no data for some entries resulting on sets without any data?
Thanks,
Hi Rene, thank you for the suggestions. I’ll check if any of those occur.