Hi,
I've extracted two times the AD_PERFORMANCE_REPORT, the 1) with no segmentation and the 2) with segmentation by date.
Report 1) shall give aggregated metrics about ads over all time whereas report 2) is used to analyze trends and especially the active time (status=enabled) of each ad.
It's surprising that when executing a left outer join from 1) to 2) on the ad_id that in roughly 50% of the cases an individual ad didn't enter an enabled state on any day in report 2).
If all ad states are considered (disabled/enabled/paused) the join performs as expected.
Note that there isn't a correlation between the impressions, clicks or conversions of an ad and if it is successfully joined or not.
In other words, although an ad isn't found with state=enabled in report 2) it doesn't mean that it has only very low to zero metrics - which is in my view confusing since an ad with a considerable amount of impressions must have been enabled at some point of time.
I'm considering for this experiment only AdNetworkType1=AdNetworkType2=SEARCH and AdType in [EXPANDED?_TEXT_AD, EXPANDED?_DYNAMIC_SEARCH_AD] and a few mio ads.
I've extracted report 2) only with the fields ad_id, ad_group_id, ad_state, date and have the feeling that I'm not receiving all rows which I expected to receive. Am I the first one experiencing this behavior?
Best,
Felix