Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hi all!
I have read the topic about slow excell export functionality of Qlik v11 SR2 in here : Extremely slow Excel export
which seems to be corrected in 11.00.11426 (the Update1 to SR2), according to comments.
However, I'm using QVx64 Desktop11.2 SR5 and export of ~35k rows and ~140 cols takes around 7 minutes. While exporting from QV server running the same version it takes ~20mins! The component is Straight Table with 2 calculated dimensions and expressions.
Time this takes is too much. Please, is there a solution for this version? a patch?
Thanks for your advices!
Hi,
The link you post is for 11.00 SR2 and not 11.20 SR2. Slowness can be cause by network speed and not Qlik. You can watch your document being created in the C:\ProgramData\QlikTech\QlikViewServer\Temp
Once it is created there it is pushed out to the user. I would test and see how long it takes to create the file.
Bill
Hi,
yes I am aware of version which that post deals with. And I'm saying that I'm using much newer and it takes 7 minutes for me to export that amount of data on my desktop. This is too slow.
140 columns !!! 🙂 that is a big spreadsheet.
Have you considered exporting to CSV instead to reduce filesize?
to export it to csv takes a half of xls, indeed, its ~3mins on my dev., so on client-server I expect ~10mins...
that is still too much and I think business would not want to change the format..
that is kind of the speed difference I would expect. The width of those records is the speed killer. Can it be split into a relational set of files with 20-30 columns?
it is ~2mins for 50columns in 35k rows. that is 1/3 of time that it takes to export it as a whole. so any major performance improve.
Is there a possibility that data model can affect an export speed of already loaded straight table component?
As Qlik's Anthony Deighton said at Qonnections 2014: "Every time a user exports to Excel, an angel looses a wing"... no just kidding.
But seriously, is this type of export done with high frequency by multiple users during the day, is it done after doing analysis with a bunch of selections, or could it be "prepared" for the users?
/Nicolai
I got your point and I agree. But for the people in business is often hard to change an old-fashioned way they do things...
my guess is that the number of users per day exporting will be around tens and not simultanously. And, yes, the data can be affected by multiple selections. That would be the real world.
Now for me, when testing, only one selection (client) matters. I'm not sure what you mean by prepared, however.
I need to find properties which can affect this speed / what can I do in order to make it faster..