Do you need multi-threading in the grouping of the script?

[warfollowmy_ver]

Hi, all!

Poll failed to create ... Please write your opinion..

It related to the topic, where more than two years ago hic replied Re: Sum() in the script using only one cpu core.

Let's see whether the QV community wants multithreading in the grouping in the script.

Of course, you can find a other way through the database, but this is not always convenient.

For example, we use the terminal server. And user dynamically generates a query.

Already built QVD data storage.

I can convert all such query and give user rights. Well, not really what we would not like to see this useful opportunity to QV?

P.S. I also want to add. Speed reading optimized QVD and place that takes QVD over SQL less at times it is for us the underlying factors to the vault in the QVD.

But such sadness that not multithreaded GROUP. Let us are working to make multithreaded.

[warfollowmy_ver]

Yes!

[tresesco]

This is certainly a highly desirable feature. Thanks for bringing it up once again. However, I don't think this is been left to be voted and then incorporate as a feature in Qlik. I read somewhere hic mentioning about it's possible technical impossibility. The load order does matter in the script and one thread has to wait for the other thread with another record while it comes to operations that need multiple records (like Group By) - that is what making it difficult (if not impossible) to implement multithreading in such operations.

But yes, without knowing much of other similar technologies implementing multithreading, may be in future we can see it here too. We can only be optimistic.

[warfollowmy_ver]

Thank you.

I have read that the hic that have no demand for multi-threading in the grouping.

Maybe it is, decided to try...

For example, DISTINCT in the script - multi-threaded. I attach with the file.

SQL at least twice as fast as in the grouping. But in other cases QVD lightning.

I want to hear the official answer whether the hope is, and how soon to multithreading in the grouping, or move towards a hybrid data warehouse, where we will use the direct download from the QVD and the grouping of the SQL.

[simonaubert]

Hello @tresesco  Well there is now an idea here https://ideation.qlik.com/app/?&_ga=2.16491830.1861944900.1716268834-1696948713.1645563170#/case/281... with lot of votes.

Let's be clear : it's today my main frustration point in performance.