topic Re: Creating fields based on each other - best approach in QlikView

Creating fields based on each other - best approach

Fri, 09 Dec 2016 05:56:33 GMT

Hi,

I have a rather large QVD file containing ~7 Million records.

These deal with usage data And I have a Teradata script that I need to rebuild with Qlikview logic.

In Teradata it was like

SELECT

field1,

field2,

field1+field2 newField1

newField1/field1 newField2

....

This is a simplified example, there are complex calculations in the original script.

As far as I know, I can only create newField1 in the same step:

LOAD *,

field1+field2 as newField1

FROM ....

What about newField2? Do I Need to insert the complex calculations instead using the Name:

LOAD *,

field1+field2 as newField1,

(field1+field2)/field1 as newField2

FROM ....

This will drive me mad....

Or do another load over all records to add the field? This will take a lot of time...

Thanks for any advice 🙂

shraddha_g — Fri, 09 Dec 2016 06:00:57 GMT

Try preceding load

Load *,newField1/field1 as newField2;

LOAD *,

field1+field2 as newField1

FROM ....

Fri, 09 Dec 2016 07:43:48 GMT

Alright, thanks.

Is this the fastest option, by the way? I needed 6 preceding loads overall because of all the dependencies and the load now takes about 17 minutes.

shraddha_g — Fri, 09 Dec 2016 07:59:11 GMT

Preceding load is better than Resident I guess

Fri, 09 Dec 2016 08:38:18 GMT

I'm still not satisfied with the load time. Anyhow, you're answer is correct, marked it accordingly.