topic Re: Reduce the data volume while distribution in QMC in QlikView

Reduce the data volume while distribution in QMC

dmohanty — Tue, 22 Aug 2017 17:50:16 GMT

Hi,

I have a QVW which has 36 months of data, reloads and distributes to end users. At the same time, need to create another QMC job below that which will only distribute a reduced data of 3 months to some set of users.

How to reduce the data to 3 months in QMC (once 36 months reloads is complete) and distribute?

Re: Reduce the data volume while distribution in QMC

felipedl — Tue, 22 Aug 2017 17:57:55 GMT

Hi Dicky,

You could make a copy of the QVW and edit it, so it only contains 3 (with some where clause while you load your tables) months worth of data and make the distribution of this app, thus maintaining the original file with 36 months worth of data.

Afterwards, you create a task and distribute it to the users.

Only an idea, that would be my approach to it.

Felipe.

Re: Reduce the data volume while distribution in QMC

dmohanty — Tue, 22 Aug 2017 18:16:38 GMT

Hi felipedl‌,

Thanks.

That is one option I had thought of too. But not interested to make a separate copy of the same with just 3 months. Because I have some 20 QVWs and copies will create maintenance issues.

So was eager to do while reload and distribute itself

Re: Reduce the data volume while distribution in QMC

Peter_Cammaert — Tue, 22 Aug 2017 18:29:23 GMT

Do you have a Publisher? In that case it's fairly easy to do.

Re: Reduce the data volume while distribution in QMC

dmohanty — Tue, 22 Aug 2017 18:34:58 GMT

pcammaert‌,

I am just trying to do at a publisher level itself. Just using the Calendar table fields.

Re: Reduce the data volume while distribution in QMC

Peter_Cammaert — Tue, 22 Aug 2017 18:53:38 GMT

I guess that means that you do have a Publisher license.

In that case, add a variable to your load script that contains the maximum number of months to include in your document. It should make sure that if the variable is empty or non-existent, the script uses a default period of 36 months.

The default task will not do anything special and as a result, the document with the standard name and distribution to the initial set of users will get the full 36 months of data.

Create another task (you can copy/paste the first one if you want), and now do the following:

At the bottom of the Reload tab, enter the variable name in the field 'Parameter name', and enter a value of

3;

in the 'Parameter Value' field. Don't forget to include the semicolon. This task will create a limited data copy of your document.
We would run into problems with the second copy overwriting the first one, but fortunately there is a solution to that. On the Reduce tab, specify a new file name in the field 'Reduced Document Name'. You can use something like

%SourceDocumentName%-3Months

for example
Then in the Distribute tab, specify a new list of destination users to distribute the reduced copy to.
Add any other options for the second copy and save the task.

By preference, attach the second task to the first one (upon successful completion). Do NOT run them simultaneously.

Best,

Peter

Re: Reduce the data volume while distribution in QMC

dmohanty — Wed, 23 Aug 2017 15:19:25 GMT

Hey pcammaert‌,

Thank you for a detailed elaboration with examples. This helped somewhat.

I was really NOT interested in reloading the whole document again after loading them for 36 months and again the same for 3 months, using the variable parameter option you suggested, because those are really huge models and even the 3 months reloads for 90 minutes. But trying out the same, it worked too.

So your above suggestion sparked another idea - NOT to reload, ONLY to distribute the QVW for 3 months of Data, using a Field Name and Field value in the Reduce Tab.

Created a Flag in the Load Script which holds the 3 months of data and used the same field to reduce while distributing.

Thank you again for the detailed support.

Re: Reduce the data volume while distribution in QMC

Peter_Cammaert — Wed, 23 Aug 2017 15:22:58 GMT

Ah, that's a great alternative solution, and a faster one indeed. Thanks for your feedback !