Group by in Qlik

[skiruthigadevi]

Hello Friends,

Appreciate your time. I have a fact table(weekly table with 104 weeks of data)  with below details 

Fact:

Yearweek,

Qty,

Transaction,

UPC,

customer_Id

from 201901.qvd(qvd); // I have to load all the 104 weeks(last 2 years weeks)

So in the chart, I am trying to find the Transaction which has 1 Qty, Transaction which has 2 Qty  and Transaction which has 3+ Qty with the expression below 

Count(Aggr(If(Sum({<_yearweek=[State_Current]::[_yearweek]>} Qty)=1, 1), Transaction))

Count(Aggr(If(Sum({<_yearweek=[State_Current]::[_yearweek]>} Qty)=2, 1), Transaction))

Count(Aggr(If(Sum({<_yearweek=[State_Current]::[_yearweek]>} Qty)>=3, 1), Transaction))

with the Yearweek as my dimension.

the performance is too slow in the chart when I try to select more year weeks., can we try group by in script to achieve the same?

could you kindly let me know your suggestions?

thank you!

[Brett_Bleess]

Kiruthigadevi, may need to attach a sample app on this one, so folks can check out the data model and everything...

Regards,
Brett

[skiruthigadevi]

Hello Brett,

Thank you, sure. Basically I am trying to do a basket analysis for transactions  in Qliksense.

I have my product hierarchy with 6 levels and the most granular is upc. I am using the below expression 

Count(Aggr(If(Sum(Qty)=1, 1), Transaction))

Count(Aggr(If(Sum(Qty)=2, 1), Transaction))

Count(Aggr(If(Sum(Qty)>=3, 1), Transaction))

I am looking for an alternative in the script as my data set is pretty huge

thank you!