We are creating a largish Qlik Sense application (250 End Users) that shares a common data set but contains an identifier that splits the data into 25 segments. A subset of users (average number of users per data segment =10) can only access their particular data segment.
There is an internal discussion re the use of section access (I’m not a fan personally) and the creation of multiple applications where each data segment is loaded using “where identifier = X”.
The use of multiple applications runs the risk of any changes having to be replicated 25 times so I was hoping to get a view on the general pros and cons of each approach and any other ideas how to best approach this type of challenge.
Just duplicating an app with appropriate adjusted datasets is the cheapest solution. If there are just a few ones and it's rather seldom/unlikely that there are much changes on them - it's a quite practically way.
But I think by 25 applications it's a rather bad idea - the needed overhead to update and maintain them especially by a complete new application is probably much higher than the benefits - also I assume that there are some disadvantages in regard to reload-times and RAM+CPU consumption.
Thanks for your response. For clarity it would be a single application duplicated 25 times, with a specific application containing the data appropriate to the end users. Application 1 would look exactly like Application 25 except for "where indentifier =X"
From this perspective I'd assumed that the RAM performance is likely to be better as rather than a single large data set being loaded every time that any user logs on, only the data set pertaining to the segment would be loaded when a user logs on. Obviously if there is heavy and concurrent usage the entire data set would loaded anyway. I'm assuming section access will load the entire data set and then apply rules but I could be completely wrong here?
From a section access perspective there are a couple of issues:
1. Lack of experience in using section access in a production environment
2. Perceived risk of error - the data segments are confidential to the users who can access them and any mistakes on our part in the section access process would expose confidential business information to non-authorised users.
I have no own experience with the section access in Sense because we use only View and there are some differences between them. But I doubt that multiple small applications will cause a lesser CPU+RAM consumption as a single big one - both by updating them and by using them through the users.
Both methods will have their benefits and disadvantages depending on the real usage in certain time-frames.
To the section access - of course you will need some time to learn it and get experienced but this could happens with uncritically applications and/or dummy applications/data - just to ensure that the logics work.
Personally I see no greater risks in using section access than in splitting the apps - by both approaches might mistakes lead to a non-authorised data access. This means both ways need to be checked carefully to ensure that they work like intended.
