Hi Adam,

Consider setting up a rule in your pipeline, so that it only executes if you merge to a particular branch.  I know your use case seems to be "PR approval" not merge, but the below rule will only execute on merges to main.

There are other free variables in the pipeline to maybe accomplish what you want.
 

  rules:
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event" && $CI_MERGE_REQUEST_TARGET_BRANCH_NAME == "main"'`

This is a very good post on the subject that is exhaustive you may want to checkout:

https://community.intersystems.com/post/continuous-delivery-your-intersy...

variables:
  SESSION_NAME:
    value: "IRIS"
  NAMESPACE:
    value: "ACME"

stages:
  - deploy

deploy main:
  stage: deploy
  tags:
    - dev
  before_script:
    - export WORKING_DIR=$(pwd)
  script:
    - echo "Initializing ACME CI/CD Build DEV..."
    - echo "Deploying code ${CI_MERGE_REQUEST_SOURCE_BRANCH_NAME} branch to instance..."
    - |
      /usr/bin/irissession ${SESSION_NAME} <<EODEPLOY
      ${IrisUser}
      ${IrisPass}
      zn "${NAMESPACE}"
      w ##class(%SYSTEM.OBJ).LoadDir("${WORKING_DIR}/src","ck",,1)
      set sc = ##class(Ens.Director).UpdateProduction()
      EOFDEPLOY
  rules:
    # only runs if merged to main
    - if: '$CI_COMMIT_BRANCH == "main"'

I hit this too...

Instead of a query, try a dbtable with a query wrapper to a temp_table...

.option("dbtable", "(SELECT name,category,review_point FROM SQLUser.scotch_reviews) AS temp_table;") 

check this post for full in and out with jdbc to databricks.

https://community.intersystems.com/post/databricks-station-intersystems-...

Thanks Alexander, this is good feedback/info, glad I poked the bear.