Data Manager is offering more and more really cool features which makes data preparation for reports highly efficient. I appreciate the direction Qliksense is going to.
But there is one issue which has sometimes a big negative impact on the usability as well as stability of QlikSense. So far I could not find any solution to limit the number of records when adding a data source. This is not by means of having a permanent filter (for the final query in the report), but temporary --> only for the time as long as I do the data preparation stuff.
This is not a new request and there have been some discussions in the past as well (e.g.: Limit data when loading in Data Manager).
My problem is: any table is scanned first before presenting the screen to identify, prepare, model and save table associations. But there is no way to define a dedicated limit for data preparation only. So whenever a table is added, a full table scan is automatically triggered. But this is a real show stopper having huge tables with million of records and many many attributes.
However when I run into big tables either it takes really lots of time to get the scan done or - even worse - it makes QlikSense abort. Not good at all.
With the latest versions I work more and more with Data Manager because it is a time issue as I have to do lots of rapid prototyping of reports. Because Data Load Editor does not support me in that way and on top I cannot mix up both for one report (sync problems), I have to stick to Data Manager.
However, what is your experience ? I cannot imagine that there are only few people with this problem.
And how to convince QlikSense to add some parameters to have a better control ?
Finally: For me the perfect solution would be to have the possibility to define with each "add table" a record limit if required. But this limit should be only valid and automatically applied when I am in the data preparation phase within Data Manager.
As soon as I want to analyse the data within the report, then the full load shall be triggered (and for sure this will be faster as the scan of data for data preparation is obsolet then).
In fact a type of different data extract management if data are scanned for data preparation support or if data are loaded for the real reporting thing.