Since June 16th, ‘Other’ has been appearing in the list of custom fields 1,2 and 3 when it shouldn’t. For example the following image shows breaking down an ‘equip weapon’ event by custom field 2, the in-game code has not been changes since launch, however on June 16th the ‘Other’ custom field has popped up, seemingly hiding our data behind it.
It could worth noting that custom field 1 has 210 fields (one for each country code)
Additionally, events that only have a custom field 1, also show Other in custom fields 2 and 3
We noticed on June 16th the web request to fetch custom event data on the creator dashboard changed, where the namespace changed from ‘CustomEventsRocube’ to ‘recommendedEvents.custom.dimensions’ and the ‘metric’ value that gets sent up was changed for each metric (eg. ‘CountCustomEvent’ is now ‘CustomEventCount’) we believe this internal change might’ve contributed to this issue with ‘Other’ appearing.
Expected behavior
No code was changed in-game, the data shouldn’t have ‘Other’ suddenly appearing on June 16th
Page URL: https://create.roblox.com/dashboard/creations/experiences/UNIVERSE_ID/analytics/custom?filter_CustomEventName=EVENT_NAME&breakdown=CustomField2
A private message is associated with this bug report
3 Likes
Hi:
Could you tell me which experience is it? So I can better help debugging.
Hey @friedrich1717, this information is in the private message, I’ve also added a direct link and an example event name, let me know if there’s anything else I can do to help.
I’m seeing this behaviour across multiple experiences since the same date; plus minor discrepancies between numbers I exported before the 16th and looking at them numbers again when I export today.
1 Like
Hi Nicholas.
We just did some data improvement so the other grouping issue should be fixed now.
Regarding the minor data discrepancy, that is just due to the daily roll up after we publish hourly data more than 7 days. That is to help reduce storage cost on our side and provide faster UI response when showing data with a longer time range.
If you want to export historical data for overall analysis, I think downloading data more than 7 days ago can give you a more accurate result at daily aggregated level. Though the difference should be very minimal.
1 Like
