Do not input private or sensitive data. View Qlik Privacy & Cookie Policy.
Skip to main content

Announcements
Qlik GA: Multivariate Time Series in Qlik Predict: Get Details
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

Load file in prod having 5GB

Hello ,

Here i have one doubt , can any one help me . If fie (txt)size in 5 GB and i want to load it in production .

System RAM is only one GB.

Should it work ?  If itdoes n't work do we have any work arround to load data.

Many thanks in adavnce .

6 Replies
matt_crowther
Specialist
Specialist

Sounds unlikely to me; 1GB of RAM is very small for a Qlikview system.

The only caveat to that is if you're only taking part of the 5GB txt file; one or two columns say, Qlikview does ustilise column compression but trying to get 5GB of data onto a 1GB machine (that has to run the OS etc as well) doesn't sound possible to me.

Hope that's of use,

Matt - Visual Analytics Ltd

New Qlikview Design Blog: qvdesign.wordpress.com

stephencredmond
Partner - Specialist II
Partner - Specialist II

Possible, but unlikely.  It really depends on the data in the file - the uniqueness of it.

There is a "rule of thumb" that talks about 10% of the original data size.  I have seen it do better, but not always.  You might get it into 500Mb.  However, during the load process it will need more memory than that and to add any charts, etc., you will need more memory.

The problem will come at about 7/800Mb when it tries to allocate the next piece of memory and then it will fail.

Never would I describe a QlikView server with 1GB as "Production".  You need to get more RAM - are you even on a 64bit OS?

Regards,

Stephen

matt_crowther
Specialist
Specialist

Stephen,

When you mention the often Qliktech quoted '10% rule' do you find it to be that is takes up 10% (ie 500mb) on the disk as a .qvw or 500mb in memory - or both? As from my experience a .qvw usually tends to take up much more RAM than it does disk space - depending on how the User Preferences > Preferred Save Format is set and the structure of the data.

I may be wrong but a .qvd is essentially a data dump of Qlikview memory; as it would be in an open Qlikview file thus allowing speedy load times whereas the .qvw is usually heavily compressed version of the data; hence we see that the sum of a .qvw's exported .qvd files uses more disk space than the .qvw itself. As I find this to be the usual case I often use a .qvw for archiving instead of the actual .qvd.

Those are my observations but as I say; I may well be wrong.

All the best,

Matt - Visual Analytics Ltd

New Qlikview Design Blog: qvdesign.wordpress.com

stephencredmond
Partner - Specialist II
Partner - Specialist II

Hi Matt,

The disk size is, as you say, completely irrelevant to the size of data in memory.  As such, it would be pretty pointless to worry about the data 10% referring to disk size.  My experience is that it refers to memory.

However, you still need memory for all the other stuff like caching of user objects.

Not sure what your point is about QVD vs. QVW and archiving, doesn't really seem relevant to the discussion.

Regards,

Stephen

matt_crowther
Specialist
Specialist

Stephen,

My simple point is that Qlikview stores data in it's various formats in different ways depending on the file type and location (disk / RAM) and this has an impact on system specification - especially when trying to work with 5GB in a 1GB environment.

I agree maybe getting a little off topic but its something I've found useful and was never explained to me by Qliktech, I had to find out through trial and "why the hell has all my 2GB of RAM been eaten by that 750mb (on disk) .qvw" style errors.

All the best,

Matt - Visual Analytics Ltd

hugo_andrade
Partner - Specialist
Partner - Specialist

About QVW (applications):
In my experience, it's not possible to work with the whole set of data with this environment.

The file size in disk is always lower than the memory usage, because QlikView compresses the data to store the file on the disk. That's why the 750mb application eats 2Gb of ram. We have a 4Gb application (in disk) that demands 3x of ram.

About the TXT:

the best thing to do is a test. Load the data and monitor the memory usage. I do think the server can't handle the whole data.

I hope that helps.

Hugo Andrade - IPC Global

Live and Breathe Qlik & AWS.
Follow me on my LinkedIn | Know IPC Global at ipc-global.com