topic Re: Returning Vs New User Login in QlikView

Returning Vs New User Login

Fri, 26 May 2017 16:04:01 GMT

Hi QV Experts,

In the dashboard attached we can see that some users have only logged in one time and some users multiple times in a day and some users have logged in several times in different days.

Based on the given data how can we show returning Vs new users per day(date) in a bar chart?

swuehl‌

stalwar1‌

Re: Returning Vs New User Login

sunny_talwar — Fri, 26 May 2017 16:11:45 GMT

Do you have QV12? If you do then try this:

Dimension

Date

Expressions

=Sum(Aggr(If(Len(Trim(Above(User))) > 0, 1, 0), User, (Date, (NUMERIC))))

=Sum(Aggr(If(Len(Trim(Above(User))) = 0, 1, 0), User, (Date, (NUMERIC))))

Re: Returning Vs New User Login

Fri, 26 May 2017 16:21:12 GMT

Thanks for the instant reply stalwar1‌ but unfortunately I am on version 11.2 and in this version your suggested expression gives me allocated memory exceeded error. Is there a workaround for version 11.2?

Earlier I had tried below for returning and new respectively but this gave me the same error as well.

If(Sum( Aggr( Count(DISTINCT [Date]), User) )>1)

If(Sum( Aggr( Count(DISTINCT [Date]), User) )=1)

Any workaround for version 11.2?

Re: Returning Vs New User Login

sunny_talwar — Fri, 26 May 2017 16:22:23 GMT

You can try just this

=Sum(Aggr(If(Len(Trim(Above(User))) > 0, 1, 0), User, Date))

=Sum(Aggr(If(Len(Trim(Above(User))) = 0, 1, 0), User, Date))

But then make sure your Date is sorted in ascending order in the script

Re: Returning Vs New User Login

sunny_talwar — Fri, 26 May 2017 16:25:31 GMT

Another alternative can be a script based solution where we create flags and use those flags in set analysis

Re: Returning Vs New User Login

Fri, 26 May 2017 17:05:49 GMT

Thank you stalwar1‌. The suggested solution works. Appreciate your quick turnaround.