topic Data Modelling - Doubts ; Required Guidance in App Development

Data Modelling - Doubts ; Required Guidance

Ranjanac — Fri, 15 Nov 2024 21:15:44 GMT

Hi All,

I am new to qliksesnse, below are the list of questions, that i have doubts after trying. Could you please guide.

1. let suppose we've multiple files, having duplicate values, now i want to load all the files into one.
Can we use link table for that ? or, is there any other way ? Attaching QVF for the same .

2. I have 2 tables having the same structure. Now i want to remove the duplicate values and load both the tables into one. How to do that ? Attaching QVF for the same .

3. When I have multiple Fact tables in QlikView, it can be handled in 2 ways, by using
concatenate or by using Link tables. Can you please help me to understand with an example.

Thanks & Regards,

Ranjana

Re: Data Modelling - Doubts ; Required Guidance

brunobertels — Thu, 09 Nov 2023 12:16:41 GMT

For 1 :

try this with JOIN

Sales:
LOAD * INLINE [
StoreID, ProductID, Sales, BudgetQty, BudgetValue
1, 1, 5, 90%, 50
1, 2, 6, 50%, 47
2, 1, 5, 95%, 41
2, 2, 4, 20%, 27
];

Profit:
join LOAD * INLINE [
StoreID, ProductID, Profit, BudgetQty, BudgetValue
1, 1, 5, 90%, 50
1, 2, 6, 50%, 47
2, 1, 5, 95%, 41
2, 2, 4, 20%, 27
];

Budget:
join LOAD * INLINE [
StoreID, ProductID, Budget%, BudgetQty, BudgetValue
1, 1, 5, 90%, 50
1, 2, 6, 50%, 47
2, 1, 5, 95%, 41
2, 2, 4, 20%, 27
];

Re: Data Modelling - Doubts ; Required Guidance

marcus_sommer — Thu, 09 Nov 2023 12:50:04 GMT

Everything in Qlik should start with a star-scheme data-model which means having a single (vertically and/or horizontally) merged fact-table (field-names and data-structures as harmonized as possible) and n surrounding dimension-tables.

There may a lot of challenges to match everything but all this work needs to be done independently of the data-model. In the end there may scenarios in which it's suitable to extend the star-scheme or very rarely to replace it with another data-model but nothing is so simple and fast developed as a star-scheme to validate data and logic and creating the first views.

Beside this are duplicates not mandatory an error else it could be valide data. Removing duplicates might a simple load distinct are doing. For a bit more complex scenarios you will need any unique identifier to flag them or removing them - very powerful would be exists() in this matter.

I suggest you just starts with playing with some sub-sets and concatenating the facts - and then step by step applying all harmonizing/cleaning/preparing tasks.