topic Incremental Load Scenario:Monthly Stacked QVDs in QlikView

Incremental Load Scenario:Monthly Stacked QVDs

Mon, 30 Jan 2012 02:30:43 GMT

Hi,

I just wanna ask how to perform this kind of scenario in Incremental Load:

Attached is the one that I followed in creating Incremental Load (Thanks to Mr.

Rob Wunderlich

and his QV cookbook)

Incremental Load Scenario:Monthly Stacked QVDs

Mon, 30 Jan 2012 05:11:30 GMT

up!

Incremental Load Scenario:Monthly Stacked QVDs

Mon, 30 Jan 2012 09:59:41 GMT

There's not really a remarkable difference when loading the the separate QVDs into the QVW over loading a huge single one. When it is about grabbing SQL data, incremental load it is a large time saver. Of course its much more time saving to load the last day or month and store it into a QVD file than it would be loading the whole data table on every reload - assuming a large data table in sql

Incremental Load Scenario:Monthly Stacked QVDs

Mon, 30 Jan 2012 13:01:14 GMT

Sir, your answer didn't help me at all..

Incremental Load Scenario:Monthly Stacked QVDs

Mon, 30 Jan 2012 13:32:40 GMT

Oh sorry, i completely misread your question and thought you'd ask for the performance.

I checked your demo file and it seems correct except the inner join statement in the Incremental Load tab and the table tab seems to be futile.

Incremental Load Scenario:Monthly Stacked QVDs

Tue, 31 Jan 2012 00:07:06 GMT

why is it incorrect? the inner joins there handles deleted data. the table that i loaded there is just for example purposes.

Thanks!

Incremental Load Scenario:Monthly Stacked QVDs

Tue, 31 Jan 2012 08:31:14 GMT

Ah ok, because of my own practice i did not expect a sql table where data gets completely deleted. We got state flags. In your case this is of course correct.

If you want to store the QVDs per month you have to split the loaded data. If not just ignore the below code

MonthTable:

load year(dss_update_time)&num(Month(dss_update_time),00) as UserMonth,

sum(1) as FlagField

resident $(vTableName)

group by year(dss_update_time)&num(Month(dss_update_time),00);

for i=0 to NoOfRows('UserMonth')-1

let vUserMonth = peek('UserMonth', $(i), UserMonth);

UserPerMonth:

Noconcatenate load *

resident $(vTableName)

where year(dss_update_time)&num(Month(dss_update_time),00) = '$(vUserMonth)'

;

if(FileSize('$(vQvdPath)' & '$(vQvdName)'&'$(vUserMonth)') > 0, -1, 0);

Concatenate(UserPerMonth)

LOAD *

From $(vQvdPath)$(vQvdName)$(vUserMonth) (qvd)

Where not Exists($(vPK));

---Join Statement---

Store UserPerMonth into $(vQvdPath)$(vQvdName)$(vUserMonth);

drop table UserPerMonth;

next;

drop table $(vTableName);

drop table MonthTable;

exit script;

please disregard any syntax errors