JPA Query Connector Setup

Add the connector where the generated query step runs:

<dependency>
    <groupId>org.pipelineframework</groupId>
    <artifactId>query-jpa-connector</artifactId>
    <version>${pipelineframework.version}</version>
</dependency>

The connector uses the application's Quarkus datasource, ORM configuration, and JPA entities. It remains separate from the Persistence Plugin, even when both features use the same database.

YAML

Declare the query once under queries, then reference it from a kind: query step.

queries:
  latest-active-risk:
    connector: "jpa"
    input: "com.example.CustomerRiskLookup"
    output: "com.example.CustomerRiskFacts"
    version: "v2"
    jpa:
      entity: "com.example.CustomerRiskEntity"
      where:
        customerId: "input.customerId"
        status:
          eq: "ACTIVE"
        score:
          gte: "input.minimumScore"
        deletedAt:
          isNull: true
      orderBy:
        updatedAt: "desc"
      limit: 1
      projection:
        customerId: "customerId"
        riskBand: "riskBand"
        score: "score"
        accountStatus: "account.status"
      result: "single"

steps:
  - name: "Load Customer Risk Facts"
    kind: "query"
    cardinality: "ONE_TO_ONE"
    query: "latest-active-risk"
    input: "com.example.CustomerRiskLookup"
    output: "com.example.CustomerRiskFacts"
    capture:
      keyFields: ["customerId"]

  - name: "Assess Customer Risk"
    service: "com.example.risk.AssessCustomerRiskService"
    cardinality: "ONE_TO_ONE"
    input: "com.example.CustomerRiskFacts"
    output: "com.example.RiskDecision"

version participates in captured query identity. Bump it when the read meaning changes, such as a predicate or projection change that should not reuse earlier captured facts.

This first-party captured JPA form remains supported and, in this release, continues to use its existing top-level queries definition. Provider-backed Query execution is not yet enabled in the runtime; do not select JPA queries through connector operation/using bindings yet.

Java records

Use small input and output records that make the read boundary obvious:

package com.example;

public record CustomerRiskLookup(String customerId, int minimumScore) {
}

package com.example;

public record CustomerRiskFacts(
    String customerId,
    String riskBand,
    int score,
    String accountStatus
) {
}

The output projection supports Java records. The projection keys are output record component names. The projection values are entity property paths.