Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hi All,
facing very strange issue but was seriouly struggling with this issue.
we have 2 phases in the project i mean we have 2 Qlik apps, phase1 and phase2.
phase1 app contains the data till 2015 jan to 2017 Aug.
in Phase 2 app we have to show the same data as phase1 along with 2 new columns, that 2 columns are coming from the data from sep 2017 onwards.
in order to match the New qvds, we are adding the dummy column (with default value) to all the Phase1 Qvds(Fact Qvds)
one new column is coming from Fact table and other new column from Dimension table.
we are converting the all old qvds(Phase1 Qvds) into new qvds with addition of new column(dummy value).
after loading the new qvds into phase2 Qlik app, it has to perfectly match with the phase1 till Aug as we are loading the same Qvds.
but in my case they are not matching with the phase1 app...
please suggest some work arounds to solve this, alone in the team, Please help me. 
was found that some thing wrong with the converted qvds which is causing the issue. by taking diff combinations like below
| Fact Qvds | Dim Qvds | Result /matching status with phase1 app |
|---|---|---|
| Phase2 Qvds | Phase 2 Qvds | Not Matching |
| Phase2 Qvds | Phase 1 Qvds | Not Matching |
| Phase 1 Qvds | Phase 2 Qvds | Matching |
with the above test we found that phase2 (converted qvds) are creating problem, but that conversion Qlik app is fine and same way i did for other apps it was perfect.
thanks stefan for your reply,
phase1 Qlik app and phase2 Qlik app both are exactly same excpet 2 filters which we added as filters in phase 2.
- how you have compared the numbers (based on what measures / fields)? - they are same apps (with same exp and charts), just the only diff is new columns as filters in the UI, and they are mentioned in Backend script.
- how you have transformed your QVDs : using an another qlik app especially for this conversion , like loading the phase1 qvd and adding a new column like 'UNKNOWN' as TYPE. and storing them in the phase2 qvd path.
- if you have compared your QVDs with other means (e.g. QViewer QVD Viewer or a record based source comparison)
we dont have QViewer, was just checking the size of the Qvd, it is matching.
and just now i took one single Qvd from phase1 and the record count is 4121844.
and after adding the dummy column ('UNKNOWN' as TYPE), i mean the count of phase2 Qvd is also same i.e 4121844.
bit surprised to see this , that how after adding a new column as well the record count is same.
For attachments please check in the main question above.
highlighted the counts difference in phase2 app.
while preparing the phase2 app , i took a copy of phase1 and added 2 new columns in the backend, accordingly i did the changes in the phase2 extractor as well to get the new columns from sep 2017 to dec 2017.
where as for all old qvds (from 2015 jan to 2017 aug) have used the simple conversion technique.
based on more analysis observed couple of things today :
1. for few tables i can see more dim values in phase2.
ex: phase1 tables contains 3 categeries 100 value is distributed among those ,where as for phase2 i can see some additional categories where 110 devide across them....
but when i check both phase1 and phase2 dim qvds they are having same number columns.
the above is repeated in couple of tables...
2. the attached pics show bit less difference but without selecting snapshot date if select same cohort month and cohort year in both apps, --- the difference is more --- no idea why we are getting the difference-- suspecting some thing worng with the these cohort calender fields..
in the attached images when i select same snapshot date from both apps without selecting any cohort date fields we can see the only difference with cohortMO -- the same issue repeated almost in all tabs..
now i dont have any idea , how i can compare the both apps and how i can prove that both the apps are showing the same data.. 
