Data USQL (No live) vs Data Explore ,with the same timeframe , don't present same information

[PierreGutierrez]

I have a question about Data Explorer, I am using the metric of "User count - estimated (by Os, app version). And I have also used USQL to obtain the number of users. For both cases in the Yape application and for a single day December 15.
But the results are different. Does anyone know what the information in the data explorer is based on? Shouldn't the information in this case match the information from USQL and Data Explorer?

[uros_djukic1 · Accepted solution]

Hi @PierreGutierrez ,

USQL is based on ElasticSearch aggregation and taken account only end user's session. 
Data Explorer is based on Cassandra (TimeSeries) , so persistent data. 
As a result of, there exists little gap or shift.

Make sure you use comparable USQL and Explorer metrics. 

For the user session count you can use:or this one if you want all users, including anonymous users:

SELECT Count(*) FROM usersession

For user count use the following for tagged users:

SELECT Count(DISTINCT(userid)) FROM usersession

More appropriate would be this metric in Data Explorer builtin:apps.other.sessionCount.osAndGeo.  versus USQL Count(*) FROM usersession
Built-in metrics | Dynatrace Documentation

In other hands, you got this warning message in Data Explorer : 

For the dimension 'App Version' the number of dimension values was reduced to its TopX of 9

which tells the metric has a top x defined and the warning indicates that if some dimensions are missing. As consequence, only the topmost app versions will be used for aggregation...
This could be truncated your data too.

Hope this helps,
Uros

[michael_2_kumme]

Why can't there be a single source of data?  The question I get from people using the dashboards is this - the Data Explorer data is different from USQL data, why?  and which one is correct?  Understanding there is an explanation, the difference in results creates some doubt in the data or the ability to accurately report metrics.  

[PierreGutierrez · Accepted solution]

Hi @michael_2_kumme 
I understand that at the time dynatrace decided not to have a single data source, because the processing of the information was different (data persistence, processing) and it was a strategic decision in:

USQL: ElasticSearch
Data Explorer: Cassandra

I think that the information in both cases is correct, depending on the scenario you are analyzing you can choose one of them.
For Example , when analyzing metrics such as "Number of users", which is a metric that you could find from both cases (USQL or Data Explorer), the result should be similar or close.
Taking into account that in USQL extrapolation may exist depending on the case.

Likewise, I understand that currently with "Dynatrate Grail", all this will change and now it is thought and planned to be a data storage technology that centralizes all the unique information.😁