Aggregated vs Transactional Data

[#1]

Hello,

Just trying to get viewpoints on why transactional data is preferred over pre-aggregated data in qlikview, my boss is suggesting using pre-aggregated sql views for our qlikview apps wrapped up weekly , quarterly,  instead of using a calendar table.  Why would this be a bad/good idea?  Thanks

[#2]

My main guess would be that it would require multiple data sets, each with their own table.  Which woul then render the start model unusable, and decrease speed a good bit, IDK

[MK_QSL]

If you want transaction by

Hour, Day, Week, Month, Quarter, Year.....

It is not a good idea to create Aggregated Data with all above time span.

In case, if you are in need of some specific transaction, you can obviously go with aggregation.

[#4]

As usual, it depends. From my experience, data should be loaded with aggregation to the smallest level which will be used as a dimension. For example: if your analysis will drill-down to a daily basis, there's no reason to load hourly aggregated values. Unless that data would be used somewhere else.