dbt to JUnit XML

Convert dbt's target/run_results.json into a JUnit XML report so Azure DevOps (ADO) can display dbt test results in the Tests tab.

This tool is designed for CI: it reads the run results produced by dbt build, generates a single JUnit report, and writes it to a file that can be published by ADO.

What it reads

• Input: dbt run_results.json (typically target/run_results.json)
• Source of truth: the results list inside that file
• Filtering: by default, only dbt tests are included (unique_id starts with test.)

What it writes

• Output: a JUnit XML file (default: dbt-junit.xml)
• Structure: one <testsuite> containing one <testcase> per dbt test

Install / run (local)

If you're using this repo with uv:

uv sync
uv run dbt-junit-xml --input target/run_results.json --output dbt-junit.xml

You can also run it directly with Python:

python -m src.main --input target/run_results.json --output dbt-junit.xml

CLI options

• --input: path to run_results.json (default: target/run_results.json)
• --output: output XML path (default: dbt-junit.xml)
• --log-level: DEBUG|INFO|WARNING|ERROR (default: INFO)
• --include-models: include non-test nodes as testcases (default: off)

Help

You can also run the following command to get some help in the terminal:

uv run dbt-junit-xml --help

or

python -m src.main --help

Response:

usage: dbt-junit-xml [-h] [--input INPUT] [--output OUTPUT] [--include-models] [--log-level LOG_LEVEL]

Convert dbt run_results.json to JUnit XML.

options:
  -h, --help            show this help message and exit
  --input INPUT         Path to dbt run_results.json (default: target/run_results.json)
  --output OUTPUT       Path to write JUnit XML (default: dbt-junit.xml)
  --include-models      Include model results as testcases (default: only dbt tests).
  --log-level LOG_LEVEL
                        Logging level (DEBUG, INFO, WARNING, ERROR). Default: INFO

Exit codes

• 0: report generated and no failing dbt tests
• 1: report generated and at least one dbt test failed/errored
• 2: could not generate report (missing file, invalid JSON, unexpected format, etc.)

Azure DevOps pipeline example

Run dbt (which produces target/run_results.json), generate the JUnit XML, then publish it:

- script: |
    dbt build
    dbt-junit-xml --input target/run_results.json --output dbt-junit.xml
  displayName: "Run dbt and generate JUnit report"

- task: PublishTestResults@2
  displayName: "Publish dbt test results"
  inputs:
    testResultsFormat: "JUnit"
    testResultsFiles: "dbt-junit.xml"
    failTaskOnFailedTests: true

Notes / tips

• If your pipeline working directory is not the dbt project root, pass an explicit --input path.
• If you only want dbt tests in ADO, do not pass --include-models (default behavior already filters to tests).