topic Most recent year of data - Expression/Set analysis in QlikView

Most recent year of data - Expression/Set analysis

zagzebski — Mon, 02 Feb 2015 18:49:37 GMT

I am tring to get the right expression to get the sales of a customer from their most recent year of sales. I tried using an AGGR expression with a FirstSortedValue expression but I must be way off cuz nothing returns:

firstsortedvalue(aggr(sum(Sales),Customer), -Year) :

Data:
Customer	Year	Sales
A	2014	25
A	2014	50
A	2013	75
B	2012	100
B	2011	125
B	2011	150
C	2013	175
C	2013	200
C	2010	225
C	2009	250


DESIRED Results:
Customer	Latest Year	Total Sales
A	2014	75
B	2012	100
C	2013	375

Re: Most recent year of data - Expression/Set analysis

Oleg_Troyansky — Mon, 02 Feb 2015 23:02:51 GMT

Unfortunately, this problem cannot be solved with Set Analysis (at least, not directly). The reason being - Set Analysis condition is calculated only once per chart, and therefore it cannot be sensitive to your dimension values.

The solution can be built using a bit of scripting. You can determine in the script what year is the latest year for each Customer. Ideally, you can calculate a new field _Last_Year_Flag in your fact table that will be assigned 1 for all the transactions that belong to the last year for the given customer. With a flag like this, the Set Analysis expression becomes independent of your dimension, and therefore it becomes valid:

Dimension: Customer

Expressions:

Latest Year = Max(Year)

Sales = sum( {<_Last_Year_Flag={1}>} Sales)

cheers,

Oleg Troyansky

Check out my book QlikView Your Business - now available to pre-order on Amazon!

Re: Most recent year of data - Expression/Set analysis

sujeetsingh — Mon, 02 Feb 2015 23:17:18 GMT

Hello Zab,

I agree with Oleg and he had given you a good solution.

Well i always believe in playing with script that in UI since it minimizes the load on object and hence gives good response time .

I will recommend you the below way of implementing your need.

Well you can go for any it is your way.

PFA the sample file

Re: Most recent year of data - Expression/Set analysis

nagaiank — Mon, 02 Feb 2015 23:19:54 GMT

If you want to solve the problem with scripting only, the following script works

load * Inline [

Customer,Year,Sales

A,2014,25

A,2014,50

A,2013,75

B,2012,100

B,2011,125

B,2011,150

C,2013,175

C,2013,200

C,2010,225

C,2009,250

];

NoConcatenate LOAD Customer,Year

,Sum(Sales) as Sales1

Resident A group by Customer, Year;

Drop Table A;

NoConcatenate Load Customer

,FirstSortedValue(Year,-Year) as MaxYear

,FirstSortedValue(Sales1,-Year) as MaxSales

Resident B Group By Customer;

Drop Table B;

Re: Most recent year of data - Expression/Set analysis

zagzebski — Tue, 03 Feb 2015 19:19:47 GMT

Thanks Oleg!

If you have a minute can you explain this, not sure I totally understand : The reason being - Set Analysis condition is calculated only once per chart, and therefore it cannot be sensitive to your dimension values.

Steve

Re: Most recent year of data - Expression/Set analysis

Oleg_Troyansky — Thu, 05 Feb 2015 00:27:03 GMT

In your initial question, you wanted to calculate sales for the last year for each customer. That means that every line of the chart should be calculated with a Set Analysis condition that is dependent on the dimension value (Customer) at the same line.

Since Set Analysis condition is only calculated once, before the chart is being produced, it cannot contain any conditions that are dependent on the dimension values.

Re: Most recent year of data - Expression/Set analysis

zagzebski — Thu, 05 Feb 2015 02:08:04 GMT

Thanks!

Steve

Re: Most recent year of data - Expression/Set analysis

zagzebski — Thu, 05 Feb 2015 02:09:21 GMT

Thanks!