FAQ

July 15, 2026 · View on GitHub

Documentation Maintained Navigation

FAQ — Qlik to Power BI Migration

General

Q: What Qlik formats are supported?

  • QVF files (.qvf) — Qlik Sense application packages (ZIP-based)
  • JSON exports — Qlik Sense Engine API / metadata exports
  • Qlik load scripts — converted via qlik_script_converter.py

Q: What Power BI format is generated?

PBI Project 4.0 (.pbip + TMDL) — the modern, Git-friendly format. Open with Power BI Desktop in Developer Mode.

Q: Do I need Qlik Sense installed?

No. The migration reads QVF files directly (they are ZIP archives) and parses JSON exports without any Qlik dependencies.

Q: Do I need Power BI Desktop?

Only to open and validate the generated .pbip project. Enable Developer Mode in Options → Preview features.

Migration

Q: How do I migrate a QVF file?

python migrate.py MyApp.qvf

This runs the full 2-step pipeline: extraction → generation.

Q: Can I migrate from a JSON export?

python migrate.py export.json --output-dir output/my_project

Q: How do I reuse extracted intermediate JSON?

python migrate.py MyApp.qvf --skip-extraction --output-dir output/existing_json

Q: What are the 11 intermediate JSON files?

FileContent
app_metadata.jsonApp name, description, author, dates
datasources.jsonConnection strings, tables, columns, types
dimensions.jsonMaster dimensions (fields, labels, groupings)
measures.jsonMaster measures (expressions, labels, formats)
visualizations.jsonChart types, dimension/measure bindings
sheets.jsonSheet layouts, cell positions
variables.jsonVariables (name, definition, comment)
loadscript.jsonFull Qlik load script
associations.jsonTable associations / relationships
bookmarks.jsonBookmarks and selections
master_items.jsonMaster items (combined dim/measure refs)

Q: What does the Data Preparation Lineage section show?

It shows the preparation flow in the comparison report: Qlik load-script steps, Power Query M steps, layer classification (Bronze/Silver/Gold/Mart), purpose tags, complexity scoring, and multi-source steps such as JOIN/CONCATENATE. For JSON-based runs, the comparison report falls back to the source app JSON script field when loadscript.json is unavailable, so lineage still renders in both single-file and batch runs.

Q: Why might the lineage section still appear smaller than expected?

The report is data-dependent. Smaller apps will only show the steps present in that app. For JSON-based exports, the report now falls back to the script field in the source JSON when loadscript.json is unavailable.

Q: How do I run batch migrations across nested folders?

Use --batch-recursive with --batch when your QVF or JSON exports are spread across subdirectories. The batch runner deduplicates by stem and processes .json, .qvf, and .qvw inputs.

DAX Conversion

Q: How many Qlik functions are converted to DAX?

175+ functions across 12 categories (string, math, date, aggregation, set analysis, conditional, inter-record, type conversion, null handling, logical, security, advanced).

Q: Is Set Analysis converted?

Yes. {<Year={2024}>} → CALCULATE(..., 'Table'[Year] = 2024). Multi-field and complex modifiers are supported.

Q: How are inter-record functions handled?

Functions like Above(field, n), Below(field, n), Previous(field), and Peek(field, offset) are converted to DAX OFFSET expressions. RangeSum(Above(X, 0, RowNo())) generates a running total via CALCULATE(SUM(...), WINDOW(-INF, 0, ALLSELECTED(...))).

Q: What about Aggr() expressions?

Aggr() is decomposed into DAX iterators:

  • Aggr(Sum(X), Dim) → SUMX(VALUES('T'[Dim]), X)
  • Aggr(Count(X), Dim) → COUNTX(VALUES('T'[Dim]), 1)
  • Aggr(Avg(X), Dim) → AVERAGEX(VALUES('T'[Dim]), X)
  • Multi-dim or unrecognized inner functions fall back to ADDCOLUMNS/SUMMARIZE.

Q: Are P() and E() set analysis functions supported?

Yes (v7). P({1} Field) → ALL('T'[Field]) (possible values) and E({1} Field) → EXCEPT(ALL('T'[Field]), VALUES('T'[Field])) (excluded values).

Q: What about dollar-sign expressions?

$(=Year(Today())-1) is expanded inline: the inner Qlik expression is converted to DAX (YEAR(TODAY()) - 1). Variable references like $(vMyVar) are resolved against the variables dictionary.

Power Query M

Q: Which data sources are supported?

25 connector types — see docs/QLIK_TO_POWERQUERY_REFERENCE.md.

Q: How are QVD files handled?

QVD files have no native Power BI connector. The migration generates a CSV-based M query with a comment explaining the QVD origin. Convert QVD to CSV/Parquet before importing.

Q: Can I chain transforms?

Yes. Use inject_m_steps() or build_m_query_with_transforms() from m_query_builder.py to add 40+ transform types to any M query.

TMDL

Q: Is RLS (Row-Level Security) migrated?

Yes. Qlik Section Access is converted to TMDL roles with filterExpression using USERPRINCIPALNAME().

  • Wildcard * entries generate an RLS_AllUsers role with TRUE() filter
  • OMIT columns are annotated as OLS (Object-Level Security) migration notes
  • REDUCTION columns are parsed into per-role reduce values

Q: Are hierarchies preserved?

Yes. Qlik drill-group dimensions become TMDL hierarchies with levels.

Q: Is a Calendar table auto-generated?

Yes. Call TMDLGenerator.generate_calendar_table() to get a complete date dimension with Year, Month, Quarter, WeekNumber, DayOfWeek, etc.

Q: How are geographic columns handled?

Column names like "Country", "City", "PostalCode" get automatic dataCategory annotations for Power BI map visuals.

Troubleshooting

Q: Power BI Desktop shows "Cannot load model"

  • Ensure Developer Mode is enabled
  • Check TMDL syntax: balanced quotes, valid data types
  • Verify relationships reference existing tables/columns

Q: Measures show errors

  • Qlik expressions may use functions without DAX equivalents
  • Check the DAX conversion log for warnings
  • Review QLIK_TO_DAX_REFERENCE.md for edge cases

Q: Visuals appear empty

  • Verify that visual data bindings reference existing model columns
  • Check that table/column names match between TMDL and visual.json
  • Ensure measures are defined in the correct table

Plugins & CI/CD (v8)

Q: How do I use the --json flag?

Run python migrate.py app.qvf --json to get machine-readable JSON output. The JSON includes status, table/measure/visual counts, warnings, and duration. This is ideal for CI/CD pipelines.

Q: How do I create a custom plugin?

Create a Python class with a name attribute and implement any of the 7 hook methods (pre_extraction, post_extraction, pre_generation, post_generation, transform_dax, transform_m_query, custom_visual_mapping). Load it via --plugins module.ClassName. See docs/guides/PLUGIN_DEVELOPMENT.md for details.

Q: Can plugins modify DAX expressions after conversion?

Yes — implement transform_dax(self, formula) and return the modified formula. Multiple plugins are chained in registration order.

Q: What happens if a plugin raises an error?

The error is logged and the pipeline continues. Plugins never crash the migration.

Enterprise Features (v9)

Q: How do I generate Fabric-native artifacts?

python migrate.py app.json --output-format fabric

This generates Lakehouse delta tables, Dataflow Gen2 ingestion, PySpark ETL notebooks, a 3-stage Data Pipeline, and a DirectLake semantic model.

Q: How do I merge multiple Qlik apps?

python migrate.py --merge app1.json app2.json app3.json

The merge engine uses fingerprint-based table matching with Jaccard column overlap scoring. Matching tables are deduplicated, and thin reports reference a shared semantic model.

Q: How do I assess a portfolio of Qlik apps?

python migrate.py --assess-server exports/

Scans all JSON exports in the directory and produces a RED/YELLOW/GREEN assessment per app with complexity scores, effort estimates, and wave planning recommendations.

Q: What does the DAX optimizer do?

After generation, the DAX optimizer automatically:

  • Rewrites nested IF chains to SWITCH
  • Simplifies ISBLANK patterns to COALESCE
  • Folds constants (e.g., 1 + 2 → 3)
  • Extracts repeated sub-expressions into VARs
  • Auto-generates Time Intelligence measures (YTD, QTD, MTD) when a Calendar table is detected

Q: Is PII detection available?

Yes. governance.py scans column names and expressions for PII patterns (email, SSN, phone, credit card, etc.) and flags them in the migration report.

Q: How does schema drift detection work?

schema_drift.py compares two versions of intermediate JSON files and reports added, removed, renamed, and type-changed columns. Useful for incremental migration scenarios.

Q: What security validations are performed?

security_validator.py checks for:

  • Path traversal attacks (e.g., ../../etc/passwd)
  • ZIP slip vulnerabilities in QVF extraction
  • XXE injection in XML content
  • Overly long file paths

Q: What monitoring/observability is available?

monitoring.py exports migration metrics to Azure Monitor, Prometheus, or JSON format. sla_tracker.py tracks per-app migration time and fidelity against configurable SLA thresholds.

Q: Can I deploy a shared model with thin reports?

Yes. After --merge, use the bundle deployer:

from powerbi_import.deploy.bundle_deployer import BundleDeployer
deployer = BundleDeployer(workspace_id="...", token="...")
deployer.deploy_bundle("output/merged/")

This deploys the shared semantic model first, then all thin reports referencing it.