PI System Connector should be able to delete values on the destination server if they are deleted on the source server
Sometimes when performing analysis recalculations, values are dropped from when they were initially recalculated. For example, when an analysis runs initially, it may calculate 15 values, which are then written to the source and transferred over to the destination as expected. However, if I perform a recalculation where I choose to permanently delete existing values and recalculate due to some of the analysis inputs having changed, and only 10 values are now written to the source server, these values are written to the destination, but the existing values that are now "incorrect" are still present. I would like for the System Connector to see that the events have been deleted in this instance on the source server and delete the same events on the destination.
Charles Rapp
shared this idea
Planning is underway to enhance PI System Connector to support global configuration enabling the propagation of deleted values from source to destination PI System.
4 comments
-
FranzK
commented
How about the state of this request? This request might help to ensure data quality. As deleting data should be handled carefully, i agree with Todd to make it configureable at least at the connector level (maybe even more granular e.g. based on the local point source or other point attributes).
-
mbailey
commented
I would assume most companies have a source and destination PI collective. I feel like I'm always asking for the same thing. We need them in sync. If our analytics are running on the source system (and AF) and we change an analytic calculation, the data is only corrected on the source side. The destination side is now wrong and has bad data! This leaves us a mess to try to clean up! When we change an analytic on the source side, and recalculate and replace the data the same needs to occur on the destination.
-
Thomas McCarthy
commented
Manual deletion of the calculated results on the destination system is a good workaround for small numbers of tags, but is impractical for larger analyses
-
Ability to have deletions replicated across systems as well (might need to be a config option on how to handle).
