Skip to main content
Announcements
Have questions about Qlik Connect? Join us live on April 10th, at 11 AM ET: SIGN UP NOW
cancel
Showing results for 
Search instead for 
Did you mean: 
alextomlins
Contributor III
Contributor III

Why shouldn't I just Join everything?

Hi Guys,

I am still learning Qlik and I understand about associations but i wondered why we choose to use associations over joins. I was having issues with circular references between a fact table and a reference table and so I just joined the reference table to the fact table.

My guess is computational power as the join is made upon the selection as opposed to going through every column/row in a wide table. I suppose also if there are one to many relationships or multiple fact tables?

Everything is working fine but i wondered what is best practice here and are there any documents available about Qlik data models and the shape/reasons why choose to do joins/lookups/applymaps/Keeps etc.

So back to my initial questions - why shouldn't i just create a wide table and if not when and why do i choose to use linked associations instead.

Thanks Qlikers

1 Solution

Accepted Solutions
avkeep01
Partner - Specialist
Partner - Specialist

Hi Alex,

For example: if you have sales per customer and you have a customer address, customer contact person, a phone number, a visit location, a country etc. You don't want them to join to your sales. In that case for every single sales order you'll duplicate all those dimensions. So you'll get way to much data which can make Qlik slower. One link between tables does the same thing as joining the data. Only it is way more efficient, because you'll need to load the data once.

check this post for the best practices:

Best Practices for Data Modelling

View solution in original post

2 Replies
avkeep01
Partner - Specialist
Partner - Specialist

Hi Alex,

For example: if you have sales per customer and you have a customer address, customer contact person, a phone number, a visit location, a country etc. You don't want them to join to your sales. In that case for every single sales order you'll duplicate all those dimensions. So you'll get way to much data which can make Qlik slower. One link between tables does the same thing as joining the data. Only it is way more efficient, because you'll need to load the data once.

check this post for the best practices:

Best Practices for Data Modelling