> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pidgeon.health/llms.txt
> Use this file to discover all available pages before exploring further.

# Generate Realistic Test Data

> Go from zero to a full test dataset in under 5 minutes — HL7, FHIR, and NCPDP messages that match your vendor's real patterns.

## Prerequisites

* Pidgeon CLI installed (`dotnet tool install --global Pidgeon.CLI --version 0.1.0-beta.1`)
* Optionally: a vendor profile from real sample messages

## Generate your first messages

Start with a basic HL7 ADT admission message:

```bash theme={null}
pidgeon generate ADT^A01 --count 5
```

This produces 5 synthetic ADT^A01 messages with realistic patient demographics, encounter data, and properly structured segments (MSH, EVN, PID, PV1).

<Tip>Use `--seed 42` for reproducible output — the same seed always generates the same messages.</Tip>

## Multi-standard generation

Pidgeon auto-detects the standard from the message type. You can also specify it explicitly:

<CodeGroup>
  ```bash HL7 v2 theme={null}
  pidgeon generate ADT^A01 --count 3
  pidgeon generate ORU^R01 --count 3
  pidgeon generate ORM^O01 --count 3
  ```

  ```bash FHIR R4 theme={null}
  pidgeon generate Patient --count 3
  pidgeon generate Bundle --count 1
  ```

  ```bash NCPDP SCRIPT (requires the licensed ncpdp-script package) theme={null}
  pidgeon generate NewRx --count 3
  pidgeon generate RxFill --count 3
  ```
</CodeGroup>

<Note>HL7 v2 and FHIR R4 generation are free. NCPDP SCRIPT is closed by default — install the member-supplied standards package first (`pidgeon data install ncpdp-script --accept-license`, NCPDP membership required).</Note>

## Use vendor profiles for realistic patterns

Real-world messages vary by EHR vendor. Use vendor profiles to match specific patterns:

<Steps>
  <Step title="Analyze sample messages">
    If you have real messages (de-identified), analyze them to create a profile:

    ```bash theme={null}
    pidgeon config analyze --samples ./epic-er-samples --save epic_er
    ```
  </Step>

  <Step title="Generate with the profile">
    ```bash theme={null}
    pidgeon generate ADT^A01 --vendor epic_er --count 10
    ```

    Generated messages now match the field population patterns, segment ordering, and value sets from your samples.
  </Step>

  <Step title="Validate against the profile">
    ```bash theme={null}
    pidgeon validate --file generated.hl7 --profile epic_er
    ```
  </Step>
</Steps>

## Batch generation with output files

Save generated messages to files for use in integration testing:

```bash theme={null}
# Single file with all messages
pidgeon generate ADT^A01 --count 100 --output ./test-data/adt-batch.hl7

# JSON format for FHIR
pidgeon generate Patient --count 50 --format json --output ./test-data/patients.json

# NDJSON for streaming
pidgeon generate ORU^R01 --count 200 --format ndjson --output ./test-data/labs.ndjson
```

## Common message type combinations

Here are typical combinations for integration testing scenarios:

| Scenario     | Message Types             | Description                   |
| ------------ | ------------------------- | ----------------------------- |
| ED Visit     | ADT^A01, ORU^R01, ADT^A03 | Admit, lab results, discharge |
| Pharmacy     | ORM^O01, RDE^O11          | Order, dispense               |
| Lab Order    | ORM^O01, ORU^R01          | Order, result                 |
| Immunization | VXU^V04                   | Vaccination record            |
| Billing      | DFT^P03                   | Charge posting                |

## Next steps

* [Validate your generated messages](/post/validation) to catch structural issues
* [Build multi-step scenarios](/guides/build-test-scenarios) with the Workflow Wizard
* [Set up vendor profiles](/post/vendor-profiles) for environment-specific testing