mismatch between raw data and QV output

[#1]

Hi All,

I am facing a very weird problem. I found that many numbers in my QV file are not matching with the actual raw numbers that are linked to the QV file. What is so wierd is that this discrepency in number is coming even when i don't use any functions(say sum, max, etc).

Most of the numbers in my QV are off by say around  +- 20 compared to the actual raw numbers.

Example: I have visits in my raw data as 334,392,096 but QV is showing it as 334,392,088. Same is happening for almost all metrics ( but not for all metrics).

I have not faced any such error earlier when I developed few dashboards. Is it because of some setting these numbers are going off?  I am very sure that I have applied all the filter conditions properly.

Thanks,

Sandipan

[#2]

Hi,

I feel, someware you have used "Supress  When value is Null" for any one of your dimension. Otherwise, try to use the object "Table Box" for all the fields which you would like to compare and export it to XL and do the cross check. May be it may help you.

[#3]

Hi Venugopal,

Thanks for your quick reponse.

I used table box and cross checked the numbers already. For some reason the numbers are not matching with raw data. I am also aware of the suppress options. I have not done any thing like that.

Moreover as I said this mismatch is occuring even when i am trying to get the numbers plainly without any function. Hence there is no fear of supress conditons also.

I think it is difficult to share the problem properly here. Let me know if its possible to call you on lync or skype so that I can share my screen with you and get your help asap.  Its quite urgent.

My email ID: sandipan.bepari@hp.com

Thanks in advance.

[#4]

Hi,

I am sorry, I don't have skype right now. Could you please try it in another way. Just take another qvw applicaiton and do simple load of only one table which you feel the problem is exist and do cross check with number of records which are loaded into qlikview.

[tresesco]

Sandipan,

Could you share your sample app?

[#6]

Hi Tresco,

Thanks for your response. Please find attached the sample app. I am also attaching the raw data file that is being fetched to get the numbers.  I have highlighted the 4 columns In raw data that are used in my sample app in yellow.

You can see that the numbers are off from the raw data even though the filters are properly used. Let me know if you need anything else.

Thanks,

Sandipan

[#7]

please find the sample raw data for the default filters of the sample app.  Due to security reasons I don't have permission to attach excel file.

Let me know if this helps.

Date	Country	Business Group	Customer Segment	Device Type	Ref	Store Catalogue Visits	Item Removal	Pre_PS_Search	Post_PS_Search
10/1/2012	GLOBAL-WW	All	All	All	All	532787872	529922816	429691904	685602688

[#8]

Hi Tresco,

Did you get any chance to look into it?

Thanks,

Sandipan

[tresesco]

Yeah, I had been going through it. But nothing such issues found. Let's try to figure it out step by step.

Step1:

First load a single field where you see the discrepancies, say Pre_PS_Search alone and take a note of data.

Step2:

Load with the entire SQL part and note the same field data.

Step3:

Load the entire script(with second table)  and note the data.

Then please let know your observation.

[#10]

As you suggested I loaded only the required fields from edit script. It is making no dofference.