topic Re: Questions about Limitations in QlikView

Questions about Limitations

Anonymous — Wed, 25 Nov 2020 16:16:04 GMT

I am hoping to use QlikView as a tool to present some analytical data sourced from a relational database. I understand that QlikView is designed to connect to many and unrelated data sources. I also understand that QlikView reads all data into its own structures for reasons of performance. However, there seem to be unnecessary hurdles for those looking to read data primarily from a well normalised relational database:

1) Why does QlikView not read the relationships directly?

2) Why does QlikView insist that related columns have the same name?

The forums seem to make it clear that I am expected to work around these issues by renaming fields and duplicating tables. For a large data model this is seems time consuming and unnecessary. Such limited expressiveness is making me wonder if I have chosen the right tool for the job. I am now wondering whether I should have started to use R and the myriad development tools available on that platform.

Advice welcome! Thanks.

Re: Questions about Limitations

jwjackso — Wed, 27 Jun 2018 18:20:32 GMT

Every tool has limitations. Check the qualify statement (https://help.qlik.com/en-US/sense/April2018/Subsystems/Hub/Content/Scripting/ScriptRegularStatements/Qualify.htm)

Using "qualify *" will prefix all fields with their table name. However, you are going to want to join the tables on key fields (use rename) so that the associative engine can show relationships that the user did not know existed.

If you use R and want the charts/lists/prompts to be interactive, you will need to combine with something like the "R Shiny" studio.

Re: Questions about Limitations

Anonymous — Thu, 28 Jun 2018 11:19:33 GMT

Thank you for the reply and the link to the Qualify statement. I still don't feel I have an understanding of why relationships cannot be read from a relational database or why only fields with the same name can be linked in QlikView.

You suggest that it's "so that the associative engine can show relationships that the user did not know existed", but I would have thought that this would lead to more confusion than benefit. For example, "name" might be a common field in many tables, although these fields are unlikely to be related. In which case, using Qualify will avoid this false assumption, but then that effectively disables the associative engine. So, either QlikView will make entirely inappropriate associations, or no associations, due to the false premise that user-assigned names are the best method of association. However, when reading from a well designed database, all of the precise relational information is available! It will have been carefully crafted into the data model and will be available for any application to read. It baffles me that QlikView has been designed to ignore this information!