Documents about Qlik Sense.
i really appreciate the work you are all doing @ Qlik
Unfourtunatly in case of the "data manager" i am wondering why follwing aspects are working weired or are
1) Why is it taking so long to add "real" (not just an excel with 100 rows) data?
In this step the user just adds data and does not make a load, so for example an database table should only be quick analyzed and just the first rows should be loaded. But in our cases it adds and adds so that 100.000 rows need some minutes. You can imagine that this is not really the thought of ad-hoc analyses, if an user waits for every added source minutes. We even tried QVD files, but their adding speed is similar slow. (Maybe a preload on server side would be useful for common data sources?)
2) Inside the Data Load Editor all Tables with the same column name would be joined, inside the Data Manger when adding two tables consisting the same column name, QlikSense adds a Prefix (table name) in front of the column and does NOT connect them. This is really annoying as we can not deliver a data model to our ad-hoc user. If we click on "Associations" there is nothing. Just after we edit the added tables, while removing the prefix in each table, the association works.
3) Why Useres can not associate added data by their own? It is really necessary that a non technical user can define visually how two data objects are connected. For such a use case their should be a GUI where the user picks the relevant columns he want to join.
4) If two tables are loaded and there is a connection, it would be nice if Users would directly get an overview which tables are associated (See picture).
Can you please explain why 1&2 are happening? Maybe we are doing something wrong?
Would be really great if you could consider 3&4 for further releases!
Hello Paul, I will have a Product Manager look at this and reply.
when the user does Add data, the data is being loaded into memory to a "scratch" app that is used to profile the data. You can think about the "scratch" app as a "playground" where the data is available to the user to data transformations, do and/or correct associations. So Add data is not only adding data, it is loading the data in memory.
You are correct, the data manager qualifies the fields. The reason for this is to avoid false positives for the association recommendation algorithm. We are working on improving the association recommendation algorithm, so that if the data is the same with a %, it would appear as a recommendation. We know currently there is an issue if the number of values in a table is a small percent of another table, the recommendation algorithm does not pick it up (the cases where there is a lookup table and fact table).
In addition to this improvement, we are working on a solution where we will enable the user to define custom associations in a visual way. So for the cases where the recommendation algorithm does not show any associations, the user will be able to define them. The same visual representation will also help the user understand which tables are associated. More is coming for the data manager, stay tuned!
Please let us know if you have any other questions.
great to hear you are working on further improvements and thanks for explaining,
but now i am a little bit confused
Why the user needs to load data after he added data, what in other words means he loaded the data already.The data of the "scratch" app you mentioned could be used so a final load would not be necessary? The actual process is not really efficient as a user waits for an analysis with a couple of big tables nearly an hour just for adding. After that, he waits the same time for loading.
Also we do not really understand why adding a QVD file (on server) is not faster than adding a relational database table (inside the network)? If the bottleneck is the analyzing algorithm maybe a server service that is analyzing corporate data sources at night would be useful, so users can add data sources already analyzed really fast.
how do I perform the auto generate date dimensions on Qlik sense web? it doesnt flow the same and isnt generating the date dimensions..
I assume you mean by web Qlik Sense Cloud / Qlik Cloud or Qlik Sense Enterprise? - starting with version 2.2 you need to go into the Data Manager and edit the table that has the date fields:
Then load the data:
As long as there is a date field - it should then create the default date dimensions as shown below.
If this does not work for you - open a new discussion in the Qlik Sense Cloud forum or other appropriate forum
Let me know.
I guess after I upgrade the Qlik version it'll work.. i'll wait and see
I have a question.
When I use the functionallity Add Data from Qlik Sense Server, have the new data available for all users or only for my session/user?
Sorry for my english 🙂