Category Archives: Tools

Deconstructing Visualization Performance

Have you ever had a Qlik Sense Visualization take longer to calculate than you like?  You may have been measuring the response time with QSDA (or the older QVF QS Document Analyzer) ,  Add Sense or a stopwatch.

Your chart is likely made up of several expressions — Measures, color expressions, maybe reference lines.  Which expression(s) are the hogs? Some? All?

In QSDA Pro version 1.4 I’ve introduced a “Viz Deconstruction” feature that calculates each chart expression individually and lets you see where time is spent.  In the QSDA Pro Viz page, click the  button under a Viz:

Clicking the  button will pop up the Viz Deconstruction dialog.  Each expression in the chart will be executed and reported individually.

For this chart I can see that one Measure takes much longer than any other expression.  Now I know where to focus my efforts if I want to improve response time.

Here’s another example. Big chart, lots of data, seven Dimensions.

The Measures require over seven seconds each to generate  7M+ output rows. Maybe less detail would result in better performance and a more usable chart.  I’ll uncheck a few Dimensions and 

Wow! Big improvement in calc time and a more manageable number of rows.

As a final example we’ll see that long calc times are not always due to Measures.

The Measures take only a few milliseconds to calculate.  But the color expression takes more then 1/2 second.  Can we do this coloring more efficiently?

You may already be deconstructing charts by making clones, commenting code and remeasuring.   Of course, you have to deal with the effects of caching and cleaning up after yourself.  QSDA Pro provides a much faster and more structured approach.

QSDA Pro is free to try with applications containing 100 or fewer objects.  If you want remove the limits, get support or acknowledge the value you get from QSDA, purchase a Monthly or Annual subscription.

I hope you find the Viz Deconstruction feature useful. I have!

-Rob

Note: There was a bug in the 1.4.0 version that kept the Deconstruction feature from calculating when using a DESKTOP Connection. This has been fixed in version 1.4.2.

 

Share

QSDA Pro as a Quality Tool

I my last two posts I introduced QSDA Pro and the flag workflow.  In this post I’m going to show how you can use QSDA Pro to quickly uncover quality issues in your Qlik Sense App.

QSDA groups flags by category, and the “Quality” category is where I usually begin my app review.  Quality flags are used to indicate where something appears to be seriously broken such as:

  • An expression syntax error.
  • A reference to a Master Measure that no longer exists in the library.
  • A missing extension.
  • A bookmarked field that is no longer exists in the data model.

These are the kinds of things you want to discover before your users do!

It’s relatively easy to create new errors in an existing Qlik app.  Qlik does not warn or block you from deleting a resource like a Field or Measure that is in use.

A properly done field rename may update names in expressions.  But it will not update variables or bookmarks.  And I’ve seen plenty of field renames go wrong and break expressions as well.

It can be difficult to detect all app errors by visually checking your app.  The broken piece may be a color expression or a calculation condition that isn’t readily visible.

So yes, easy to break, sometimes hard to detect and heros find  problems before their users do.  Enter QSDA Pro.

In QSDA menu,  Help -> Flag List will display a list of all potential flags in the installed version. Here’s the current list for Quality (The “{}” bits are values that will get filled in when the flag is created.)

Some of these are fairly straightforward, some a bit more subtle.

Why would you have a missing Master Dimension or Measure?  A fairly common scenario is not understanding that Master Items are referenced by internal Id, not by Name.  So you create something called “Sales” and use it in several charts.  Later you are doing some exploration and create something called “Sales2” which you like better. So you think if you delete “Sales” and rename “Sales2” to “Sales” everything will be using the new “Sales”.  Wrong.

Because the master item is now gone, I can’t tell you what used to be in this master item.  You will have to rely on an app backup. Or…if you have a previous QSDA Analysis for this app you can look up the master item definition there!

All Expressions and Dimensions are validated using the Qlik syntax checker. A typical error might be a bad fieldname (perhaps because the field was removed from the model).

Flag Details will give us a detailed error message, location of this expression  (Master Library) and importantly the Use Count. This use count is zero so we know deleting the measure is a viable option.

QSDA uses the Qlik syntax checker so the test is only as good as what I  get from Qlik.  There are some limitations, for example when column labels are used in an expression.  You’ll note that the Qlik expression editor flags this as invalid so QSDA will pass this message on.

The “Parse Error” flag is raised when Qlik Syntax says the expression is ok but QSDA finds the expression does not conform to expected syntax rules. This could be a problem with the QSDA parser  (a work in progress) or it could be Qlik tolerating something unusual.  For example, do you think this is a valid expression?

sum({<Sales={">10 between < 90"}>}Sales)

The Qlik syntax checker will declare it valid. And it will return an official looking result — which may or may not be correct.  But what is that “between” keyword? (if you’re curious why “>10 के बीच < 90” also seems to work, see https://masterssummit.com/something-to-get-confused-about/).

Nervous about refactoring or changing your apps?  Can you relate to “Change Paralysis”  or Fear Driven Development (FDD)?  We all need a tool like QSDA Pro to proactively plan changes as well as audit after the fact for unintended errors.

Download QSDA Pro now and analyze some of your own apps.  QSDA Pro is free to use during the beta period.

You can analyze published and unpublished apps, so go ahead and take a look at some of those production apps.   Let me know in the comments if you find something interesting!

-Rob

 

 

 

Share

QSDA Pro Flags

In my first post introducing QSDA  Pro, I told how Document Analyzer was a tool I developed for use in my own consulting work and shared with the community of Qlik developers.  The Qlik Sense QVF version got a big boost in utility when Axis Group joined the project and contributed their many many years of Qlik experience.

In this post I will  introduce the  Flag workflow of QSDA.

QSDA communicates it’s advice through “Flags”. Flags are issues that warrant your attention.

The Summary page gives an overview of Flags for this analysis. Clicking any of the Flag boxes or Flags in the Nav Ribbon will take you the Flags page.

The Flags page provides a collapsible list of all flags and buttons for more detail and flag status.

 

Flag text identifies the error and in most cases provides some detail, such as a fieldname or number of bytes that could be saved.

The dimension failed validation. The error is: Bad field name(s) "[PostalCode]" .

Click  to  show the Flag Details panel.

The details panel will provide more context for the flag to help you evaluate your response.  Depending on the flag, there may be additional hyperlinks.  Click the “?” icon to link to the Help article for this flag type.

To support a workflow of “checking off” the flags,  buttons and flag status indicators are available in the both the Flag Details and Flag List views.

 Mark this flag instance status  as “Resolved”.

 Mark this flag instance status  as “Hidden”.

Resolved and Hidden flags are not shown in the list unless the filter is activated on the filter bar. Flags with no status are “Unresolved” and are always shown.

Buttons will automatically toggle to allow you undo a status. You can set/unset statuses on individual flags or at the flag type level.  Hover over a button to discover its function.

From the Flag List you can view all flags.  In the resource pages, you will also see the same flags but attached to the resource.  For example, on the Dimension page,  any associated flag will be shown with the specific Dimension.

These workflow features allow you to punch through the list of issues in your app, hiding those you don’t care about and checking off issues as you resolve them.

Learn more about QSDA Pro and download the latest version.

-Rob

 

 

 

Share

Introducing QSDA Pro

QSDA Pro is the next generation of my Document Analyzer tool for Qlik Sense.  I published the first Document Analyzer for QlikView in 2009 followed by the Qlik Sense version in 2017.  I think it’s fair to say both are the most widely used Qlik developer add-on tool.

Qlik Sense Document Analyzer (QSDA) has been extremely useful. But it’s had some limitations, especially in the server environment. I’ve wanted to address those shortcomings,  and now I bring you my response :

QSDA Pro is a completely rewritten standalone app that provides  features of the current QVF based QSDA and more.  QSDA Pro can analyze and provide advice on apps in QS on Windows, Qlik Cloud, and QS Desktop.

QSDA can help you produce cleaner, smaller and faster apps by identifying:

  • Unused resources, such as fields, that can be removed.
  • Errors in your app such as missing fields, syntax errors or missing key values.
  • Bad practices and recommendations on how to remediate.
  • Highlight performance hotspots.

QSDA Pro can do all of the above and more for public sheets and private sheets.  “Is anyone using this field in a personal chart or bookmark?” is no longer an unknowable thing.

QSDA is also your goto companion for those ad hoc tasks like modifying use of a field. Maybe a renaming or dividing by 1000.

Where am I using the field “Company Name”?  Easy.  Filter on the field name and I can instantly see the field is not referenced by Variables, but it is used in 22 Charts, 23 Dimensions and 5 Expressions. Oh and 1 Bookmark.

What Charts where?  Click Viz on the ribbon to get a list.  I can explore the chart Dimensions and Expressions there or click a link to open the associated sheet in the Qlik Sense Hub.

QSDA is smart! It can determine that all four of these measures calculate the same thing, even with differences of variables, casing and whitespace.  And it recommends you should settle on one Measure while very helpfully showing how many times each variation is used to help you in your decision.

In the coming weeks I’ll be posting more about the capabilities of QSDA Pro.  I’ll also be responding to feedback on the QSDA Pro Community site.

QSDA Pro is currently in public beta testing and is free to use during the testing period, which is expected to last for several months.

You can learn more about QSDA Pro and download your own copy.

Licensing terms for the supported production version will be announced later. There will be a free tier which may have  limits on volume and  features.

-Rob

 

Share

QViewer + Easymorph, Better Together

Thousands of Qlik Developers use my QViewer QVD Viewer tool every day to browse inside QVD files.  Sometimes you want to do more than browse. I get questions like

  • Can I sort the rows?
  • I want to sum field x grouped by field y. Can I do that?
  • Can I get a histogram of field values?

The answer to all of the above — and more — is Yes!  When you install the free Easymorph Data Prep and ETL tool, you can use the powerful features of Easymorph like an extension to QViewer.  With Easymorph you can sort, query, transform and profile your QVD data.

When viewing a QVD, press the button on the menu bar. The current QVD will open seamlessly in Easymorph.  If Easymorph is not yet installed you will be taken to a webpage with download and install instructions.

You can also open Easymorph by right-clicking a QVD from QViewer folder view.

The QVD will open in Easymorph and you can then use all of Easymorph’s powerful features with your QVD.

Easymorph has plenty of help and an active Community to help you get started.  I’ve also prepared a couple of Help articles in QViewer showing some typical QViewer/Easymorph  tasks.

-Rob

 

 

Share

CubeTester

When working on Qlik Sense performance issues I frequently find I want to measure the performance of specific expressions. I might want to know how variations of an expression may perform against each other.  In a slow chart with many measures I want calculation time individually for each measure to focus my efforts.  Or perhaps I’m just satisfying a general curiosity or trying to settle a bet.

You can measure the performance of expression variations by modifying the chart and measuring the overall chart response time with something like Chrome Add Sense or QS Document Analyzer.  That approach can get kind of clunky especially when you are focused on a subset of measures in the chart.

I prefer a more structured approach to testing expressions. The tool I reach for is CubeTester.

CubeTester is an open source Nodejs command line tool for testing the performance of Qlik HyperCubes (Dimensions and Measures).  The test specification is written in a json file as either a HyperCubeDef or the “simplified” Dimension/Measure syntax.

Here’s a sample test written in simplified syntax that tests three variations of a cube (chart) containing one Dimension and three Measures.

I’ll run  CubeTester specifying the file that holds this test:

node index.js test tests/columns.json

And receive this output:

There is no significant difference in performance between the variations. Importantly, I can also see that all three return identical  total values as well.

CubeTester supports two commands:

  • test : Run tests.
  • extract: Extract app charts into a test file.

There are a number of options that can be specified on the command line or in the test definition. See the readme for more information on available options.

in addition to testing variations or trying out a theory, here are some other cases where I’ve used CubeTester.

  • When working with a mashup where my HyperCube exists only in code, there is no chart to test.
  • In a slow rendering chart I can test individual measures, combinations of measures and non-data expressions (like color expressions) to find the culprit.

Using CubeTester I can quickly try out ideas and document my progress as I work through an issue. I’ve made some interesting discoveries!

Some notes:

  • Testing against a server uses certificates for authentication.  (Pull request welcome if you want more auth options).
  • Make sure you specify “wss” when using a server endpoint eg
    wss://your.server:4747
  • You’ll need to test with enough data to get calculation times of sufficient magnitude.  Two results of 5 milliseconds vs 7 milliseconds are not precise enough to draw conclusions from.
  • Calculation time is affected by the capacity of the target machine and what else is running.  I recommend to repeat tests until you see a stable pattern.  Use the –repeat option and take the lowest result from each repeat.

CubeTester is free to use. Have fun!

-Rob

 

Share

Data Browser Tricks

Summary: I demonstrate my latest “Data Browser” sheet for use in Qlik data modeling. Download here.

Today I want to share the latest version of my “Data Browser” sheet.

What’s a “Data Browser”? It’s my name for something you may already have. The ubiquitous “System Sheet” or sometimes “Profiler” where you may have listboxes or charts utilizing the system fields “$Field”, “$Table”, “$Rows”.  These fields are part of the shadow data model that hold useful metadata about the data loaded in your app.

Some very useful example profile sheets have been shared in the community over the years.

You can download the latest QlikView and Qlik Sense versions of my Data Browser here.

For QV, I just copy the objects into my new app in one go. For QS,  I either copy the charts one-by-one or if I’m lucky they are already in my template (who’s going to build the sheet copy extension?) .

I’ll do this walk through with the QV version, the QS version is similar.

When I make selections in $Field I can see values and frequency counts in a listbox. After selecting for example the “Sales” field, the values and frequency counts are shown in a listbox. I also have a  histogram and some descriptive statistics about the field values.

I can select field values and drill into data using  this sheet or my application sheets. For example, I might want to select the outlier high Sales value, then select Product in $Field to find out what Product is associated with this Sale, clear the Sales field to see what other Sales look like for this product and so on.

There are several properties such as Null Count (Information Density), field alpha/num content ($tags) that I can get from the built in Table Model Viewer.  The Viewer requires me to examine one field at a time. In the chart below I get an overview of those properties of interest.  Because it’s a chart, I can use sorting and selecting to focus.

Something I don’t get from the Table Viewer is the difference between numeric count and text count for a field.  I like to surface this problem (highlighted in yellow) early in my modeling.

One of my favorite features is the “Value Association” chart below. It’s likely a favorite because it took me a long time to work out the expression!

In  Table Viewer we use “Subset Ratio” to understand where we have connected and un-connected data in our model.  Subset Ratio is limited to reporting the relationship between key fields. It can’t tell me how data field  values in this table associate to data field values in other tables.  Subset ratio, like other stats in the Viewer, does not respect selections. For example, if I select a specific Customer, how many SalesReps are linked to the Customer?

The chart below (highlights added) covers all these use cases and can also surface problems in the model.

I’ll start by selecting a field central to my model, in this case “OrderID”.   I then sort by the “Pct” column.  This column represents the percentage of field values associated to the selected field,
“OrderID”.

In the green highlight, I’ve called out a Table “Sales” that has zero association with Orders. Something to look into.

In the orange callout, I can see that only 19% of my Employees are associated with Orders.  That might be a candidate for trimming the Employees table when loading.

In the red callout, I see something puzzling. Only 47% of my OrderDates are linked to Orders. That may be ok, I would need to review the data.  What looks not ok is that WeekDay Pct is also 47%. I would have expected something like 5/7 or 6/7. No fraction of weekdays would equal “47%”.  And there are 1092 values for WeekDay…something is off.

I’ll select WeekDay from the chart and examine the values in the listbox. Aha, the Weekday values were created incorrectly.  They were created with a Date() function instead of the correct WeekDay() function.

I used to hide the sheet before production but now I generally leave it in as I and others find it useful in production as well.

Besides the Table Viewer, another Qlik modeling tool I really like is Catwalk. If you haven’t used Catwalk I encourage you to check it out.  I won’t go in to explaining Catwalk as it has an excellent tutorial built in.

Do you have any favorite profiling or modeling tools to use with Qlik?

-Rob

Share

Form Input and Commenting in Qlik

A common desire from Qlik customers over the years has been the ability to interactively edit data on a QV sheet and persist the changes in Qlik or maybe write back to another system. As Martin Mahler of VizLib describes it “turning a one-way data street into a two-way street”.

Sometimes the term “writeback” is used to describe the idea of inputting data in a Qlik app and persisting that data — writing back  the data — to some other system.  The new data is saved and shared with other users of the app. Importantly,  in a form/writeback application, the added data is associated with some row of the data model. For example, in a warranty claims app we may have a dropdown that allows the user to categorize each claim.  The assigned category is used in further analysis.

I think of “commenting” as making annotations on a chart, or chart data point, and a given set of selections.  Ideally, comments may turn into a discussion and use some sort of notification mechanism.

These are not pure terms. There are overlaps to be sure.

I’ve seen some interesting bespoke implementations by partners.  There is also a growing list of off the shelf products that enable writeback and or commentary within Qlik Sense or QlikView.

Some products are Qlik Sense only, some work with both QlikView and Qlik Sense.

Most products allow you to create table sheet objects that mix Qlik DImensions and Expressions with additional input fields such as freeform text, dropdowns or checkboxes.  They all persist the data to some type of backend store such as a database.

When evaluating a product for your requirements, here are some items to consider:

  • Do you have requirements for read-only and update users?
  • Are you looking to add additional data in a single chart or do you need to reference the new data from multiple charts?
  • Are you looking to add one to one new data or build complex workflow apps?
  • Does your business objective require structured form data,  free-form commentary or both?

There are an interesting range of products and capabilities out there.  Klikins and Emark Forms for example let you add new fields to a straight table.  One of the more interesting approaches is K4 Analytics, which embeds Excel into Qlik. This provides the full range of Excel formula and formatting functions. You can build some pretty powerful aps this way.

There are products that focus on finance reports. TrueChart creates a set of functional and beautiful IBCS compliant reports along with a nice navigation interface that can be reused throughout the Qlik app.

Both TrueChart and Climber Finance Report support the type of commenting and annotations user require in finance reporting.  I’m excited that the Climber commenting is being expanded and released as a generic commenting & collaboration product by VizLib. Qommentary is another global commenting solution.

Here in no particular order are some Writeback / Commenting products I’m aware of. The headings are links to the vendor site.

I’m sure I’ve missed some products.  Feel free to leave me a comment if you have something to share.  It would also be good to hear about use cases where you have found value in implementing an input/writeback solution.

Inform Write 

QlikView and Qlik Sense

K4 Analytics

QlikView and Qlik Sense

TrueChart

Qlik Sense

Emark Forms

Qlik Sense

Klikins

QlikView and Qlik Sense

Pomerol Writeback

Qlik Sense

Qommentary

Qlik Sense

VizLib 

Qlik Sense

Komment

Qlik Sense

Share

Qlik Sense Document Analyzer V1.5

Version 1.5 of the Qlik Sense Document Analyzer (QSDA) tool is now available for download here.

If you are not familiar with QSDA, it’s a free application profiling tool for Qlik Sense that can help you identify items such as unused fields, poorly performing expressions and data model problems.

Here are the significant changes in V1.5:

The Installer allows editing of all install paths and installation will create a log file in the User’s temp directory.

The  Summary ribbon at top of each sheet provides for sheet navigation by clicking a cell.

 

Improved error handling in the connector. Certain types of errors, such as a chart calculation timeout, will not terminate the script. Error count is reported on Summary sheet and detailed error messages can be viewed on the “Extract Log” sheet.

Items that should be considered incomplete due to analysis errors will highlighted in yellow.  This highlighting is a work-in-progress as I discover new possible errors.

Field widths (Symbol width) are no longer estimated! Field sizes are obtained directly from the engine.  Due to this change, this version will report different values for field sizes than previous releases, generally +/- 10%.

The connector edit dialog has been improved.  Error messages now appear properly in their own window and a progress bar displays when the list of applications is fetching.

If you have general usage questions on QSDA, please use the comments section here or Qlik Community.

If you have a bug to report, use the issue tracker.

The only installation supported by the installer is QS Desktop. You can use QS Desktop to analyze applications on a server.  It is possible to manually install on a server by uploading the application (qvf), the qsda-ribbon extension and the connector directory if you have the need and want to give it a shot.  I do plan to create a supported server installer in the future.

Have fun!

-Rob

Share

Qlik Sense Document Analyzer v1.4

Version 1.4 of the Qlik Sense Document Analyzer (QSDA) tool is now available for download here.

If you are not familiar with QSDA, it’s a free application profiling tool for Qlik Sense that can help you identify items such as unused fields, poorly performing expressions and data model problems.

I’m pleased that Mike Steedle of Axis Group has joined me as co-developer of QSDA.   Mike contributes his many years of experience in profiling and tuning Qlik applications.

 

Here’s what’s new in version 1.4:

  1. Installer improvements. The correct directory for Connectors and Apps is automatically detected.
  2. New attribute “ObjectIsExtension”. Possible values are:
    • False – not an extension.
    • True – object is an extension but not a widget.
    • Widget – object is a widget.
    • Missing – object is an extension, but extension is not present on server.
  3. Unused Master Visualizations, and the Dimensions and Expressions therein, are now extracted and processed.
  4. DimensionLabel field added to “Dimensions” table.
  5. New table “Bookmarks”. Bookmarks are now extracted and linked to Fields.
  6. MB Constant changed from 1000*1000 to 1024*1024. This means these numbers now scaled in MB will have slightly smaller values than previous versions.
  7. Calc Time is now displayed in seconds instead of milliseconds.
  8. Dimensions and Expressions summary added to Objects sheet.
  9. Some analysis of the data model is performed, and the results and recommendations expressed as flag fields. New sheet “Flag Summary” will display an overview of found conditions.
  10. Color highlighting of detected problems.
  11. Reorganization of sheets and sheet layouts.
  12. Addition of a “Glossary” sheet provides descriptions for flags.

The installer currently only supports installation on Qlik Sense Desktop, although you can  analyze applications on Enterprise Server.

Improving the server analysis capabilities and possibly a server install will be a focus of the next update.

If you have general usage questions on QSDA, please use the comments section here or Qlik Community.

If you have a bug to report, use the issue tracker.

I hope you find QSDA useful.  I’m excited to see it maturting and pleased to have the help from Axis Group.

-Rob

 

 

Share