18 Replies Latest reply: Jul 10, 2017 11:12 AM by Weston Kingsley

# Set Analysis with Multiple ANDs

Hello Qlik Experts,

I have two (maybe three if necessary) tables I'm looking at for this one, Donors and Gifts:

Donor:

DonorID,

AssignedFlag (this is a Y or N)

DonorID

(An optional third is a Gift Calendar):

I need to Count the set of Donors who are Assigned that gave a gift in 2015, and compare that to Donors that were assigned that gave a gift in 2015 AND in 2016.

Calc 1 = Donors with a 'Y' in AssignedFlag AND who have a Gift Date in 2015

Currently I'm using this calculation:

This seems to be working correctly, but would appreciate feedback if I'm missing something...

Calc 2 = Donor with a 'Y' in AssignedFlag AND who have a Gift Date in 2015 AND who have a Gift Date in 2016

This is the Calculation I'm having trouble getting correct and am looking for help on.

Basically we're trying to calculate who gave a gift 2 years ago and of those donors, who gave again in 2016.

Thanks in advance for the help!

• ###### Re: Set Analysis with Multiple ANDs

To make things a simple as possible in the app I would create these flags in your load script.

2015Donated Flag and 2016DonatedFlag

Then in the app you don't have worry about anything but checking these 2 flags.  No complicated Set Analysis.

• ###### Re: Set Analysis with Multiple ANDs

Definitely willing to take that solution for a spin. Then a count with AND conditions will work.

However, I'm concerned this isn't an expandable solution. What happens next year when this rolls around? I guess I should've phrased this question as:

What Assigned Donors gave two years ago, then also gave again last year?

• ###### Re: Set Analysis with Multiple ANDs

Calculate variables in your script and use them instead of hard coding years.

• ###### Re: Set Analysis with Multiple ANDs

Calc 1 = Donors with a 'Y' in AssignedFlag AND who have a Gift Date in 2015

=COUNT({<DonorID = p({1<AssignedFlag = {'Y'}, GiftYear = {'\$(=Year(Today())-2)'}>}DonorID)>}DISTINCT DonorID)

Calc 2 = Donor with a 'Y' in AssignedFlag AND who have a Gift Date in 2015 AND who have a Gift Date in 2016

=COUNT({<DonorID = p({1<AssignedFlag = {'Y'}, GiftYear = {'\$(=Year(Today())-2)'}>}DonorID) * p({1<AssignedFlag = {'Y'}, GiftYear = {'\$(=Year(Today())-1)'}>}DonorID)>}DISTINCT DonorID)

• ###### Re: Set Analysis with Multiple ANDs

Thank you Manish, this seems to be working!

Can you explain the use of "p" and the "1" before each set?

• ###### Re: Set Analysis with Multiple ANDs

Set modifiers with implicit field value definitions ‒ QlikView

Try to understand the use of P() and E() from this link.

This is also called as Indirect Set Analysis.

DonorID = P(DonorID)

Mean Possible DonorID

DonorID = p({1<AssignedFlag = {'Y'}, GiftYear = {'\$(=Year(Today())-2)'}>}DonorID)

Here Possible DonorID which are having AssignedFlag = Y and GiftYear 2 years back compare to today's year

1 is used to ignore the filters users will do in application, so that the end result would not get affected.

Regards,

MK

• ###### Re: Set Analysis with Multiple ANDs

mrkachhiaimp Thanks for the links on this. This helped my understanding quite a bit after seeing it applied to this solution, so thanks!

As a follow up question, now that I have these 2 list of Donors with the qualifications, I'm trying to figure out what the Money lost is from 2015 to 2016. The difference between Calc 1 and Calc 2 is what we're calling "lapsed donors". If I subtract Calc 2 from Calc 1 I get the list of Donors who are assigned that gave in 2015 but NOT 2016. I'd like to find out how much money we lost  assuming they would've given the same amount.

Does this make sense? How would I go about calculating this?

• ###### Re: Set Analysis with Multiple ANDs

Provide sample data here.. I will try to give you expressions for that.

• ###### Re: Set Analysis with Multiple ANDs

Sure.

Donors

DonorIDAssignedFlagTotal2015Total2016

001

Y500010000
002Y100000
003N250005000
004N50000
005Y20000
006Y2000010000
007Y05000

1010011/1/2015

5000

1020011/1/201610000
1030022/1/201510000
1040033/1/201525000
1050033/1/20165000
1060044/1/20155000
1070055/1/20162000
1080066/1/201520000
1090066/1/201610000
1100077/1/20165000

With this data, the donors we would "lose" money from are 002 and 005 because they gave in 2015 but not in 2016 (Notice Donor 004 also gave in 2015, not in 2016, but isn't Assigned). The total money lost from these two Donors, if they would've given the same amount HAD they given, would be \$12,000 (\$2,000 for Donor 005 + \$10,000 from Donor 002).

Thanks for taking a look!

• ###### Re: Set Analysis with Multiple ANDs

Try this

=SUM(

AssignedFlag = {'Y'},

>}

or

=SUM(

AssignedFlag = {'Y'},

>}

• ###### Re: Set Analysis with Multiple ANDs

This worked perfectly! Thank you.

Now I'm just trying to modify it to add another calculation to add up the money lost from people who DOWNGRADED or logically, Gave Less in 2016 than 2015, but not working yet.

Based on the tables above, the result would be \$22,000. Donors 002, 005, and 006 all are assigned AND gave less in 2016 than 2015.

The greater than expressions so far haven't worked. How do I compare the two values, then sum the differences to see the money lost?

• ###### Re: Set Analysis with Multiple ANDs

I think it should be 32,000

=SUM(

AssignedFlag = {'Y'},

>}

• ###### Re: Set Analysis with Multiple ANDs

I'm still seeing \$22,000.

Donors 001 and 007 increased their giving.

Donors 003 and 004 are not assigned.

That leaves 002 (\$10,000 decrease), 005 (\$2,000 decrease), and 006 (\$10,000) . Did I miss something?

Just want to make sure the calculation is right, so if the expression you mentioned returns 32,000 then must be off, right?

• ###### Re: Set Analysis with Multiple ANDs

Check my 2nd expression, which is giving 22,000

• ###### Re: Set Analysis with Multiple ANDs

When I use the first script I get \$30,000 and when I use the second I get \$20,000. It looks like it's not including Donor 005 when I use selections.

Is there a way to do this using the Totals in the Donor table instead?

Here's the load I'm using, there could be a mistake there (I've tried it with and without the date function):

Donors:

DonorID,AssignedFlag,Total2015,Total2016

001,Y,5000,10000

002,Y,10000,0

003,N,25000,5000

004,N,5000,0

005,Y,2000,0

006,Y,20000,10000,

007,Y,0,5000

]

;

101,001,'1/1/2015',5000

102,001,'1/1/2016',10000

103,002,'2/1/2015',10000

104,003,'3/1/2015',25000

105,003,'3/1/2016',5000

106,004,'4/1/2015',5000

107,005,'5/1/2016',2000

108,006,'6/1/2015',20000

109,006,'6/1/2016',10000

110,007,'7/1/2016',5000

]

;

• ###### Re: Set Analysis with Multiple ANDs

107,005,'5/1/2016',2000

This must be 5/1/2015

Date looks wrong to me..

• ###### Re: Set Analysis with Multiple ANDs

Ah, I missed that. Numbers were running together.

Thanks! expression is verified to be working. Excellent work.

Can you explain what each expression is doing?

From what I can tell the first equation is Summing the total gift amount from 2015 where donors gave more in that year than 2016. The second looks to do the same thing, then subtracts the amount given in 2016.

Is this correct?

• ###### Re: Set Analysis with Multiple ANDs

Or use this

=SUM(

AssignedFlag = {'Y'},

>}

-

SUM(