Qlik Community

Qlik Design Blog

All about product and Qlik solutions: scripting, data modeling, visual design, extensions, best practices, etc.

Employee
Employee

It was back in April at Qonnections, our 10th global partner conference that we unveiled Qlik Sense 2.0 and shared our platform strategy with the world.  It was also the first time we talked in detail about our plans for QlikView 12. Today I’m delighted to be able to share the news that it’s arrived! QlikView 12 will you please stand up and show yourself to the world!


There is no doubt that this is an eagerly awaited release by many of our 37,000 strong global customer base. But why? QlikView is a very mature product, it's functionally rich, and it’s undoubtedly in my opinion the product that revolutionized business intelligence and ultimately created the global data discovery market as we know it today. So what is so important about QlikView 12?


An investment in QlikView 12 is an investment in Qlik


With QlikView 12, Qlik delivers on its commitment to its proven, market-leading data discovery solution which secures our customers long term investment in the product.  It also lays the foundation for our customers to partner with Qlik to build out their business intelligence strategies and meet the expanding needs of their BI consumers by addressing multiple use cases through a unique platform approach to visual analytics.


QlikView 12 now runs on the second generation QIX (Qlik Data Indexing) engine that powers the entire Qlik portfolio. With this improvement, we can more easily help customers address new use cases in Qlik Sense by allowing them to share data models across the platform. 


Our investments also benefit the way our customers use QlikView today. QlikView 12 delivers a number of deployment, performance, security and connectivity enhancements along with greater accessibility through enhanced mobile touch-enabled capabilities. In addition QlikView customers will be able to now take advantage of Qlik’s strategy to deliver value added cloud services – such as Qlik’s “Data as a Service” offering, Qlik DataMarket.


(If you want to see some of this in action check out this brief presentation)


QlikView 12 - What's New Presentation


QlikView - REST Connector

QlikView - Qlik Data Market


Put simply, QlikView is a business intelligence solution with an unrivaled pedigree, functional richness and delivers the lowest cost of ownership in the market. Many customers have already delivered robust guided analytics and dashboards to knowledge workers across their organizations, and with QlikView 12, that investment is secured. 


Regards,


Michael Tarallo

Senior Product Marketing Manager

Qlik

@mtarallo - follow me



22 Comments

Thanks for sharing the videos. QlikView 12 looks great.

I wanted to know if it is possible to install QlikView 12 without having to remove QlikView 11.20? In essence can we have QlikView 12 and QlikView 11.20 both installed on a single machine at the same time?

Best,

Sunny

473 Views
Employee
Employee

Hi Sunny - I believe the only option is to upgrade your existing installation. I would not recommended installing both versions in the same system anyway. If this is for testing and migration purposes, I would assume following a proper migration and testing plan would be in order.

I can find out more information if need be.

Regards,

Mike

473 Views

"QlikView 12 now runs on the second generation QIX (Qlik Data Indexing) engine that powers the entire Qlik portfolio. With this improvement, we can more easily help customers address new use cases in Qlik Sense by allowing them to share data models across the platform."

I am intrigued by being allowed "to share data models across the platform" and would love an example scenario to help me understand how one could exploit this.

0 Likes
473 Views
borisman
Contributor III

Bill, it works even with 11.2 - you can binary load your QVW or QVDs, created with QlikView 11  to Qlik Sense application. My bet with QV12, you can now also do a reverse thing - load data from QVFs back to QV - Mike, can you confirm please?

0 Likes
473 Views
jamesdutoit
New Contributor

We use QV macro's to create and distribute various excel reports. Does QV12 still support macro's?

Kind Regards

James

0 Likes
473 Views
Luminary
Luminary

Yes, QV12 support macros .

0 Likes
473 Views
Luminary
Luminary

There is no option to load a QVF in QV12.

0 Likes
473 Views
Not applicable

So QlikSense has responsive design. Does 12 have responsive design with the new engine? I see improvements with touch technology, but it's not clear if that means we have to design a 'tablet' version, a 'mobile version' or if we just design one report and the responsive design automatically works across those devices.

Thanks!

0 Likes
473 Views
borisman
Contributor III

it is same old QV11 - all they did they did a better handling of touch events. So no, it is not responsive design or anything.

0 Likes
473 Views
srikanthsri1
Valued Contributor

what is the main difference in term of Qlikview engine and QIX engine ?

it does same de duplication of values  logic and same associative model.?? how different is the new QIX engine in Qlikview

Regards

Raj

473 Views
Employee
Employee

Hi Raj - the QIX engine is the new term to describe - the Qlik Indexing Engine which is the core in-memory component that provides on-the-fly aggregations, compresses the data and stored and indexes the dimensional values that increases analysis performance. Overall the same experience is still achieved with the way it worked previously, including the associative experience.

Apart from a number of bug fixes and optimizations, the major improvement is our change from: Row-based storage to Columnar. The old Qlik engine (QlikView 11) has row based internal data tables, whereas the new QIX engine (Qlik Sense + QlikView 12) has columnar data tables. This change provides performance improvements, a modern storage design and data model compatibility between QlikView and Qlik Sense. For example, any data models stored in the form of .qvd files, whether created by Qlik Sense or QlikView - will work with both products.

HTH

Regards,

Mike Tarallo

Qlik

473 Views
srikanthsri1
Valued Contributor

Oh thanks for clarification Mike ...

just initially thought that QlikView engine is Columnar Architecture as in the below link

http://www.dbms2.com/2010/06/12/the-underlying-technology-of-qlikview/

Business Discovery Blog Archive: A Conversation... | Qlik Community

Hope to see many more advancement in next version from Qlik

Regards

Raj

473 Views
jonas_rezende
Valued Contributor

Mike.

The QlikView Server has also been updated to version 12?

Regards,

Jonas

0 Likes
473 Views
borisman
Contributor III

Have anyone actually tested and confirmed performance improvement in 12?

0 Likes
473 Views
borisman
Contributor III

thanks for sharing, Jonas!

0 Likes
473 Views
borisman
Contributor III

thanks Jonas for reporting back!

0 Likes
473 Views
Employee
Employee

Hi Jonas - et al - We will have a quick review of this and also update you on our findings. Initial reports have shown an increase in performance, but your feedback is very valuable and we appreciate you taking the time to perform these tests. I will update you all shortly.

Kind Regards,

Mike T

Qlik

473 Views
borisman
Contributor III

this would be great, Mike!

I did my own testing yesterday on my new laptop. I used Open HWEtest_parallel_1G.qvw app that Qlik Scalability folks normally use for application testing with JMeter (it has 1 billion rows) and I also used our most intensive app I could find (nice star schema design with 64 Million rows). First app uses very simple straight expression like SUM and COUNT. Second app uses everything - complex set analysis, AGGR etc.

In both cases performance was very similar. I repeated all tests 4-5 times:

   

Appuncompressed size on disk, Gb# of fact rows
HWEtest_parallel_1G.qvw8.71,000,000,000
myapp.qvw5.264,000,000
Test CaseQV 11.2 SR10 x64, secondsQV 12 x64, seconds
Time to open HWEtest_parallel_1G.qvw23
My app - tab 11110.6
My app - tab 220.618
My app - tab 320.618
My app - tab 45.25
473 Views
jonas_rezende
Valued Contributor

Hi, Boris.

What was the RAM consumption?

Regards,

Jonas Melo.

0 Likes
473 Views
Employee
Employee

Hello Boris an etl al - I have received the following updated information in regards to the performance improvements.

The performance improvements will vary by application and not every application will see improvements. Generally you will need to add more columns to an data model to see more improvements or if you are using a data model that is joined by association.

The performance improvements are in 2 areas:

  • Wider fact tables. It is impossible to state a fix number of columns where QV12 is faster than QV11, but adding columns will show where the breakpoint is.
  • Shared file caching, gives improved response times in QVS clusters, especially in ramp-up scenarios where large number of sessions are started in a short time-period.


Hope this is helps


Mike T

Qlik

473 Views
borisman
Contributor III

it helps, thanks Mike. second improvement sounds very interesting actually

0 Likes
473 Views
Partner
Partner

Hi Mike,

Thanks for the update. May I suggest that for performance improvement statements such as this, that you consider releasing a supplementary example benchmark app with accompanying scalability lab test results against a defined hardware spec showing the lab proven performance increase under well-defined pre-conditions? The issue with high level statements like 'more columns = greater performance improvement expected' is that it's fuzzy, imprecise and often gets perpetuated like Chinese whispers. It reminds me of the old Qlik pre-sales statement that each user should add between 10% - 15% memory consumption overhead relative to the base document memory footprint. An easy message that has been heavily circulated amongst Qlik and Qlik partners alike, but which is completely and utterly untrue! For those that are curious about the truth - users just increase volume of queries against the server, hence the speed at which the cache starts saturating with results. Leave a single user on a server for long enough and they may eat up the full lower memory band (default server setting: 70% of total) all by themselves. This is a pretty intuitive cache design and does not imply any issue with catering for large volume of users. There is of course more to it than what I've described but never the less, my explanation is infinitely more accurate than the pre-sales statement of 10% - 15% overhead per user.

Note that the application I tested performance against is exactly that - an application with wider than normal fact table, yet we have not identified any performance improvement of 12 over 11.20 SR2 (or SR13, tested today). For reference, my fact table is 77 fields wide - sparsely populated structure, mainly measures. Of course hardware and infrastructure introduces other variables that could skew results, but it seems unlikely that a like for like local test (not over network) on the same machine would not manifest any obvious performance improvement should this statement be true as currently articulated.

I appreciate that it is order of magnitudes more effort to provide hard evidence than to simply describe expected performance improvement, yet I feel this is exactly what is needed to deliver release announcements with a high degree of confidence that the improvement claims will stand up under scrutiny. It will also give the community easy, instant tools to dry run an upgrade and verify first hand if their specific environments indeed will be able to realize stated benefits.

Food for thought.

Kind regards

Jonas

0 Likes
473 Views