Qlik to Microsoft Fabric Reporting: Set Analysis, QVDs and Section Access Explained
Three Qlik building blocks and where each lands in Microsoft Fabric reporting.
Qlik to Microsoft Fabric is not a lift and shift. A Qlik app combines a load script, QVD files, an associative model, Set Analysis expressions and Section Access in one package, and each part has a different home in Fabric and Power BI. This guide explains the three that cause most rework in reporting: Set Analysis, QVDs and Section Access.
It is written for BI leaders and architects who run QlikView or Qlik Sense and are weighing a move. If you are still comparing platforms, start with our legacy BI vs Microsoft Fabric total cost comparison.
QlikView End of Life or End of Support: What It Means for Reporting
Qlik supports QlikView by release, not as one product with one end date. Qlik's lifecycle page, as quoted in search results when we checked on 18 November 2026, lists QlikView 12.80 as out of support from 31 October 2025, 12.90 out of support from 31 October 2026, and 12.100, released in September 2025, supported to 30 September 2027. We could not open the page itself, so confirm the dates on it before you quote them.
So this is a risk question, not a deadline. A release that falls out of support matters most for regulated or board-level reporting, where an unpatched platform is hard to defend. Treat the lifecycle page as one input to your Qlik to Microsoft Fabric reporting plan, alongside licence cost, skills and the number of apps people still open. Our Tableau migration playbook follows the same method for a different tool.
Qlik to Microsoft Fabric Reporting: Associative Model vs Star Schema
Qlik links tables automatically when fields share a name, and its engine treats the whole model as one connected set. Selections flow in every direction. That convenience also causes synthetic keys, which Qlik creates when two tables share two or more fields, and circular references that need a link table to fix.
Power BI works differently. Relationships are explicit, usually one to many, and filters flow in a defined direction. The cleanest target is a star schema with fact tables in the middle and dimensions around them. Qlik itself recommends star schemas, so many well-built apps already point the way. Plan to remodel rather than copy, and put shared dimensions such as date, customer and product in one model that many reports reuse.
Load Scripts to Pipelines, Dataflows and Notebooks
A Qlik load script does four jobs at once: it extracts data, transforms it, builds the model and sometimes stores intermediate files. Split those jobs in Fabric so each one is visible and reusable.
| Qlik script task | Fabric home | Note |
|---|---|---|
| Extract from databases and files | Pipelines or Dataflows Gen2 | Land raw data in a lakehouse |
| Joins, mapping loads, cleansing | Dataflows Gen2, notebooks or SQL | Choose by skills and data volume |
| Incremental loads | Pipeline or notebook merge into Delta | Keep a watermark per source |
| Model shaping and link tables | Star schema in the semantic model | Fix synthetic keys upstream |
Dataflow Gen2 can write straight to a Lakehouse destination, which gives the model one governed source. The aim is that business logic lives once, upstream, instead of inside each app.
Replacing QVDs with OneLake Delta Tables
A QVD is Qlik's own columnar file, used to stage and reuse data between apps. In Fabric the same job belongs to Delta tables in a OneLake lakehouse, which every engine and every report can read. A common layout is a bronze layer for raw data and a gold layer shaped for the model.
QVD layers are replaced by Delta tables in OneLake. Illustrative architecture.
There are two routes. The preferred one is to re-source from the original systems with pipelines or Dataflows Gen2, so you stop depending on Qlik. The second is a one-off export of history that cannot be re-sourced. Qlik's STORE statement can write Parquet, for example Store mytable into myfile.parquet (parquet);, which you can then load into Delta tables. Parquet is strongly typed, so check fields that Qlik treats as both text and number.
Choose the model mode after the tables exist. Direct Lake on OneLake reads Delta tables directly and does not fall back to DirectQuery, and V-Order is recommended for tables that Direct Lake will read. Import stays valid for sources outside OneLake. Our Microsoft Fabric capacity estimator helps size the capacity.
Set Analysis to DAX for Reporting
Set Analysis changes the selection that one Qlik expression sees. The set identifier says where to start: 1 means all records regardless of selections and $ means the current selection. Modifiers then change specific fields. DAX does the same job with CALCULATE and filter functions, because a measure runs in a filter context that you can add to, clear or replace.
| Qlik Set Analysis | DAX pattern (illustrative) | Note |
|---|---|---|
Sum({<Year={2025}>} Sales) | CALCULATE ( [Total Sales], 'Date'[Year] = 2025 ) | A set modifier becomes a column filter |
Sum({1} Sales) | CALCULATE ( [Total Sales], REMOVEFILTERS () ) | Ignores slicers, but RLS still applies |
Sum({$<Region=>} Sales) | CALCULATE ( [Total Sales], REMOVEFILTERS ( Geography[Region] ) ) | Clears the selection on one field |
Sum({<Customer=P({1<Category={'Bikes'}>} Customer)>} Sales) | CALCULATE ( [Total Sales], CALCULATETABLE ( VALUES ( Sales[CustomerKey] ), REMOVEFILTERS (), Product[Category] = "Bikes" ) ) | P() is a table of values. E() uses EXCEPT |
Avg(Aggr(Sum(Sales), Customer)) | AVERAGEX ( VALUES ( Customer[Customer] ), [Total Sales] ) | Aggr() becomes an iterator over a dimension |
Rank(Sum(Sales)) | RANKX ( ALLSELECTED ( Product[Product] ), [Total Sales] ) | Respects the user's slicers |
RangeSum(Above(Sum(Sales), 0, RowNo())) | CALCULATE ( [Total Sales], FILTER ( ALLSELECTED ( 'Date'[Date] ), 'Date'[Date] <= MAX ( 'Date'[Date] ) ) ) | Running total over the date axis |
$(vTarget) variable | A measure, or a what-if or field parameter | Variables become model objects |
These are starting patterns, not a converter. Test every translation against the output of the Qlik app, and expect Aggr() and P() or E() to need the most care.
A worked example, illustrative only. A Qlik chart shows sales for customers who bought Bikes, regardless of the category the user has selected. In Set Analysis that is a P() set on Customer with {1} as the base. In DAX you build the list of customer keys with Bikes purchases, remove other filters inside that list, and pass it to CALCULATE as in the fourth row above. The base idea carries over, and the syntax does not.
Section Access to Row-Level Security
Section Access is a security table in the load script, with fields such as ACCESS, USERID or USER.EMAIL and a reduction field. Dynamic data reduction limits the rows a user sees by matching the reduction field to a field in the data. Qlik requires field names and values in that section to be uppercase, and a reduction value that is not in the data reduces nothing unless Strict Exclusion is on.
A Section Access table rebuilt as a security table and a dynamic RLS role. Sample data only.
The Power BI equivalent is dynamic row-level security. Build a security table of users and what they may see, relate it to the data, and filter a role with USERPRINCIPALNAME(). A user who belongs to no role sees no data, which is stricter than the Qlik default.
Two checks matter. RLS applies to workspace Viewers, not to Admins, Members or Contributors, so give report consumers the Viewer role. And Test as role uses your own identity, so test dynamic RLS by signing in as real Viewer accounts. In Direct Lake, RLS defined in a semantic model applies within that model, while OneLake security applies across engines, so decide where security should live. For help with the wider design, see our data governance consulting.
- Qlik support ends by release, so check Qlik's lifecycle page for yours.
- Rebuild the associative model as a star schema, not a copy.
- Replace QVD layers with Delta tables in OneLake, ideally by re-sourcing.
- Translate Set Analysis by controlling filter context, and test every measure.
- Rebuild Section Access as dynamic RLS and test with real Viewer accounts.
App-by-App Migration Waves for Qlik Reporting
Start with usage. Count how often each app opens, who owns it, how many Set Analysis expressions it uses and whether it has Section Access. Retire what nobody opens and merge duplicates. Then work in waves.
Wave one is the shared foundation: lakehouse tables and a star schema for the busiest domain. Wave two is simple apps with few Set Analysis expressions. Wave three covers the heavy apps, where Aggr() and P() patterns sit. The last wave moves the apps with Section Access, once the security model is proven. Run Qlik and Power BI side by side through one full reporting cycle and have each business owner sign off totals before cutover. Our Qlik to Fabric migration service runs this as a structured assessment and build.
Qlik to Microsoft Fabric FAQ and Next Steps
Can Qlik apps be converted to Power BI automatically?
Not completely. Microsoft's migration guidance says conversion cannot be fully automated end to end, so treat any converter output as a draft and rebuild the model, measures and security against the original.
Does QlikView have an end-of-life date?
Qlik publishes end-of-support dates by release rather than a single product date. Check Qlik's QlikView product lifecycle page for your release before you plan.
Is Set Analysis the same as CALCULATE in DAX?
They solve the same problem, which is changing the filter context of one calculation. The syntax and the evaluation model differ, so translate the intent and test the result.
Can I move QVD files straight into OneLake?
Qlik can store tables as Parquet with the STORE statement, and Parquet can be loaded into Delta tables. For most estates, re-sourcing from the original systems is the cleaner route.
How do I test Section Access after migration?
Sign in as real Viewer accounts and compare each person's totals with the Qlik app. Test as role uses your own identity, so it is not enough for dynamic security.
Should the new model keep Qlik's associative design?
No. Power BI needs explicit relationships, so design a star schema with fact and dimension tables and use link tables only where the data needs them.
A short Qlik migration assessment confirms your app inventory, your waves and your first model. Our Qlik to Fabric migration team works alongside Power BI consulting and Microsoft Fabric migration specialists, and our guide to migrating to Microsoft Fabric covers the wider program. Speak with a certified consultant to turn this into a plan for reporting you can trust.