Skip to main content
Announcements
Introducing Qlik Answers: A plug-and-play, Generative AI powered RAG solution. READ ALL ABOUT IT!
cancel
Showing results for 
Search instead for 
Did you mean: 
sifatnabil
Specialist
Specialist

Commencing ReduceData issue

Hi all,

I have an issue with a large document (around 10 million rows in total). On my PC (with 12 GM RAM 64 bit) it works fine (slowish on calculations in objects), and 100% QVD based. It takes 1 minute to load the 10 million rows of QVDs. However, on the production QlikView server, it takes one hour on this step:

“(2015-04-28 08:52:33) Information: Commencing ReduceData”

Any ideas on how to resolve this?

1 Solution

Accepted Solutions
marcus_sommer

I don't know what these message meant but it looks that the qv server hadn't enough cpu/ram ressources (have a look in the performance settings within the qmc and in the taskmanager) or the server pulled the data over a slow network. Then normally such qvd load is very fast.

- Marcus

View solution in original post

2 Replies
marcus_sommer

I don't know what these message meant but it looks that the qv server hadn't enough cpu/ram ressources (have a look in the performance settings within the qmc and in the taskmanager) or the server pulled the data over a slow network. Then normally such qvd load is very fast.

- Marcus

sifatnabil
Specialist
Specialist
Author

Thanks Marcus, I also believe the issue was CPU usage, due to some high intensity calculations being done on objects over millions of rows (like with firstsortvalue()). My reasoning for this, is that I loaded a blank test file onto to the server with the exact same data, and it was able to load 11 million rows in 1 minute. But when done with the front-end objects, it produced that error and took 1 hour to load.

So I rebuilt the document from scratch, and optimized it such there are

  • 1 dropped table
  • Only 8 variables
  • 25 dropped fields
  • 65 less objects
  • 187 less expressions
  • RAM footprint reduced to 223 from 261 MB

And it successfully loads in 1 minute now.