Potential event data aggregation bug in 2.36.13.1

Hi all,

This morning I updated our program database from 2.36.11.1 to 2.36.13.1

Suddenly I noticed that all the integer values in our events were no longer being summed but rather counted (ie 1 count per event, regardless of the numbers in the data element).

Here is what the data should look like:

Here is what it looks like after the update:

If I check the events themselves I can see that the data is actually correct:

Anyway, I have restored last night’s backup and downgraded back to 2.36.11.1 and everything is working again.

Has this been reported yet? If not I will try to reproduce the issue and get a Jira logged.

Edit: I should add, I did all the usual incantations, such as clearing app cache, deleting and rebuilding analytics tables, restarting everything, checking that the data elements were actually set to SUM aggregation, and praying to the DHIS2 gods.

1 Like

Hi @HaydnJ

I found this jira: [DHIS2-14355] - Jira
Could you take a look at the jira, and possibly confirm this is the same issue as you’re experiencing?

Caroline

3 Likes

This does look very similar.

The event data elements I was using were Positive of Zero Integer but this seems close enough.

I will try to reproduce my issue on Play and then add my version to the issue if it is present.

2 Likes

Hi @Caroline,

I am able to reproduce the issue on the play version of 2.36.13.1 so I have added the version to the linked Jira issue.

Thanks for pointing that out to me.

1 Like

hi @HaydnJ ,

Thanks for flagging this.

There will be no more maintenance patched for 2.36, as it is now out of support, but this is quite an unfortunate regression, so we will try to get a new hotfix out to fix it as soon as possible (in the next days).

Kind regards,
Phil

2 Likes

Hi again @HaydnJ

We have now released a hotfix for this issue (see DHIS2 patch release 2.36.13 is now available - #3).

Thanks again and kind regards,
Phil

3 Likes

Thanks for the fast turnaround team!

Edit: Just installed and all working perfectly.

1 Like