topic Re: Help with Service Level in QlikView

Help with Service Level

Fri, 13 May 2016 09:34:11 GMT

Hi guys, I'm quite a newbie with Qlik and I need a little help with a typical example of a Service Level.

Service Level: I've got a database in which each row represent an order delivered to the customer. I've got two important variable:

- confirmed data;

- delivered data;

The difference: delivered data-confirmed data= number of days for the delivery (In a new column I've calculated the number of days).

For Example: Article Delivered Date Confirmed Date Number of days

x 10/01/15 15/01/15 5

Now I've classified the number of days in 4 categories ([classe LS])in the Data Load Editor:

Load *,

if([num giorni]<=0,'LS 0',

if([num giorni]<=7,'LS 7',

if([num giorni]<=15,'LS 15', 'LS >15'))) as [classe LS];

Here it's what i've made, what i would like to do is to visualize the percentage of different classes/categories per Confirmed Date or (Month/Year.. I've created the dimension Month/Year from the Delivered Date).

In the bar chart I've got a several problem with the creation of the percentage of [classe LS].

For Example:

Article Confirmed Date Classe LS

1 10/01/15 LS 7

2 10/01/15 LS 7

3 10/01/15 LS 15

4 10/01/15 LS 15

5 10/01/15 LS 15

So this could be that, for the 10/01/15, the 2/5% refers LS 7 (delivered in under 7 days) and the 3/5% refers to LS 15.

I've a problem in the implementation of this measure--> percentage of classe LS

In any case thanks for your time!

Re: Help with Service Level

swuehl — Fri, 13 May 2016 17:52:21 GMT

Maybe it's just a chart with two dimensions [Confirmed Date] (or Month / Year (could even be a dimension group)) and [Classe LS] and a singel expression

=Num(

Count([Article]) / Count(TOTAL<[Confirmed Date]> [Article])

,'#.00%','.',',')

Re: Help with Service Level

Sat, 14 May 2016 07:22:33 GMT

Thank you swuehl,I've just try the solution that you recommend me. The only problem is that in some cases the percentage are incorrect

Example

Month/Year Confirmed Date Class % in the graph %correct

jan/2014 10/01/14 LS>15 10,9% 1/4%

jan/2014 10/01/14 LS 0 32,7% 1/4%

jan/2014 10/01/14 LS 7 43,6% 1/4%

jan/2014 10/01/14 LS 15 12,7% 1/4%

Re: Help with Service Level

swuehl — Sat, 14 May 2016 11:25:18 GMT

I am not really sure what to count. Or maybe you need a Count(DISTINCT Article)?

=Num(

Count(DISTINCT [Article]) / Count(TOTAL<[Confirmed Date]> DISTINCT [Article])

,'#.00%','.',',')

If you are using more than two dimensions or a dimension group, you could add all dimensions except Class to the TOTAL field list:

=Num(

Count(DISTINCT [Article]) / Count(TOTAL<[Month/Year],[Confirmed Date]> DISTINCT [Article])

,'#.00%','.',',')

Re: Help with Service Level

Sat, 14 May 2016 12:32:35 GMT

I' m really grateful for your time.

In Excel the expression it's usually made counting the class.

I' don't need the distinct part.

Num(Count( [classe LS])/Count(Total<[classe LS],[DATA CONFERMATA]> [classe LS]),'#.00%','.',',')

With this expression the intervals are correct (precedent example) but the percentage took the 400%.

Re: Help with Service Level

swuehl — Sat, 14 May 2016 20:17:14 GMT

I don't think you should add [classe LS] to the TOTAL qualifier field list:

Num(Count( [classe LS])/Count(Total<[classe LS],[DATA CONFERMATA]> [classe LS]),'#.00%','.',',')

If this doesn't work, could you upload a small sample application or a complete description of your chart settings, dimensions and expressions?

Re: Help with Service Level

Mon, 16 May 2016 18:44:57 GMT

It was the correct . The REAL problem was that I was looking to the incorrect numbers given from a Qlik Table. It was showing 4 values but there were hidden 55 values.

Thank you very much!!

Re: Help with Service Level

swuehl — Mon, 16 May 2016 19:24:50 GMT

Ok, great!

If your request is resolved, please close this thread by flagging helpful and / or correct answers.

Qlik Community Tip: Marking Replies as Correct or Helpful