Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Do you really think the formation of a synthetic keys considered bad in Qlikview? If not please can you explain me.
Formation of synthetic keys is not bad, but, It's poor data model.Try to eliminate the synthetics keys.
In QlikView, if there are several fields in common between two or more than two tables then a table containing the set of composite keys is made. This table is identified by a synthetic key ($Synkey).
However, this approach by QlikView to manage overriding fields consider as a sign of poorly designed data model. The user must check for logical errors that in turn reflects into the script not making a good data model.
NO its not bad, though it is sign of poor data model but some times it plays an important role where you don't want data islands.
Depends on situation.
In QlikView, if there are several fields in common between two or more than two tables then a table containing the set of composite keys is made. This table is identified by a synthetic key ($Synkey).
However, this approach by QlikView to manage overriding fields consider as a sign of poorly designed data model. The user must check for logical errors that in turn reflects into the script not making a good data model.
Synthetic keys are not inherently bad -- in fact they are a good solution if you truely need a composite key between two tables -- however good data model design will in most cases eliminate the need for a composite key.
However, unintended synthetic keys are usually a sign of a scripting error (eg not dropping a temporary work table) or a poorly thought through model design. And nested synthetic keys and associative loops are always a bad thing, as it is very hard to get the expected results in such as model and performance may be very poor as well.
Below is a Design Blog post from Henric Cronstrom regarding Synthetic Keys that may be useful as well:
Regards,
Brett