Do not input private or sensitive data. View Qlik Privacy & Cookie Policy.
Skip to main content

Announcements
Join us in Bucharest on Sept 18th for Qlik's AI Reality Tour! Register Now
cancel
Showing results for 
Search instead for 
Did you mean: 
cbushey1
Creator III
Creator III

Link Table help

Hi all,

I read through the existing posts and couldnt find my solution so I thought I would ask.

I am trying to create a link table as I have a data model with 10+ tables. I started by creating my own key by concatenating some fields together in each table and pushed all the "keys" from each table into the Link Table.

Here is what I have so far:

TempLink:

Load

    (FOCUS_ID_NO & '_' & CodeBlueLocation & '_' & CODE_DATE) as %KeyCodeBlue

    ,FOCUS_ID_NO

    ,CodeBlueLocation as Location

    ,CODE_DATE as EventDate

Resident TempCodeBlue;

Concatenate(TempLink)

Load Distinct

    (INCIDENTID & '_' & [Event Date] & '_' & [Location where Event Occurred]) as %KeyRLIncident

    ,INCIDENTID

    ,[Event Date] as EventDate

    ,[Location where Event Occurred] as Location

Resident TempIncident;   

Concatenate(TempLink)

Load Distinct

    (AbstractID & '_' & Account & '_' & LocationName & '_' & EventStartDate) as %KeyMediLOS

    ,AbstractID

    ,Account

    ,LocationName as Location

    ,EventStartDate as EventDate

Resident TempMediLocations;

I have realized though that I am getting a lot of dupes for EventDate and I really dont want to keep loading the same date into that field.

Any suggestions on how to make this link table work more efficiently?

1 Reply
petter
Partner - Champion III
Partner - Champion III

Duplicates in itself is less worrying in QlikView in a single field because it internally/automatically do a de-duplication that is highly efficient.

In a concatenated keyfield it is much more problematic.

You should use the AutoNumber() function to make the key as storage-efficient as possible and make sure that the EventDate is really just a date and does not contain any time in it by for instance doing a Floor(EventDate) and also turns into an integer. This will make sure that you get the best distinct ratio for the EventDate ( unique values / all values ).