QV document eating memory

[#1]

Hi,

I have one QV document that is eating memory from my server. The thing is that when I open this document with the AJAX (and firefox) client, the server starts using 6 GB of memory, but while I'm working (switching tabs, applying filters, selecting pivot tables, ...), the memory keeps increasing to 34 GB and so, until I have to restart the server (so it is not a problem of not having enough memory. I think that the issue might be in the pivot tables I'm using and the Count expression in them. Is it possible that QVS recalculates the count every time without erasing the old ones in memory?

I use simple expressions the pivot table like:

- Count(transaction_id)

- Count(session_id)

- Calculated Dimension expression (allows me to select defined top N answers. The table is resized showing only the chosen N): =IF(aggr(rank(total aggr( Count(transaction_id),answer),0,1) ,answer)<=topAnswers,aggr(rank(total aggr( count(transaction_id),answer),0,1) ,answer))

This is my table schema

[suniljain]

what is the application size ?.

[#3]

What do you mean by application?

The document size is 285 MB and the pivot tables are between 6 million rows the bigest one and 1.5 million the smallest

[kaushalview]

Hi

Ferran

Cases of Problem:

1.If you have create expression using Aggr function then qlikview take time to calculate the value depend upon that expression

2.when QVS restart its delete the data from the busser every time (Old data).

regards

Kaushal Mehta

[suniljain]

Perform all calculation that you are doing in pivot table to script level.

it definitely improve the performance of application on server dutring application access.

[#6]

But then I will not be able to recalculate tables right? Imagine, the data of my tables is people asking questions and operators giving answers. So, I have one table containing all the questions ordered by count (say 10000 questions). Then, I want only to show top 5 questions --> my table will show only the first 5 most repeated questions. This I could do it in the script, calculate the count only once, and then just cut the table and so on. But, the thing is that I can apply filters. Let's say I have a contain filter : table, chair, computer, bed. When I choose table, now I see top N answers ordered by count containing "table", so, the first count is not valid anymore and I need to recalcute...

Did I explain myself? it can sound a little bit tricky