Define branch-specific environment variables in Jenkinsfile - jenkins

Question

I'd like to pass environment variables by publish stage based on the branch that the multi-pipeline job is processing.
While this example below works, I don't like how I'm ending up with additional stages, one per branch. Because I'm using the withCredentials plugin, I need to have the MY_CREDENTIALS variable set before that block. Is there a more elegant approach to solve this?
pipeline {
agent {
dockerfile true
}
stages {
stage("Config vars for staging") {
when {
branch 'staging'
}
environment {
MY_CREDENTIALS = 'creds-staging'
MY_ENVIRONMENT = 'production'
}
steps {
sh "echo MY_ENVIRONMENT: $MY_ENVIRONMENT"
}
}
stage("Config vars for production") {
when {
branch 'master'
}
environment {
MY_CREDENTIALS = 'creds-prod'
MY_ENVIRONMENT = 'staging'
}
steps {
sh "echo MY_ENVIRONMENT: $MY_ENVIRONMENT"
}
}
stage("Publish") {
steps {
withCredentials([[$class: 'AmazonWebServicesCredentialsBinding', credentialsId: env.MY_CREDENTIALS, accessKeyVariable: 'AWS_ACCESS_KEY_ID', secretKeyVariable: 'AWS_SECRET_ACCESS_KEY']]) {
sh 'make takecareofit'
}
}
}
}
}

Answer 1

Use switch statement. In example below I forced FAILURE if branch is not master or staging:
pipeline {
agent {
dockerfile true
}
stages {
stage("Config vars") {
steps {
script {
switch(branch) {
case 'staging':
MY_CREDENTIALS = 'creds-staging'
MY_ENVIRONMENT = 'production'
break
case "master":
MY_CREDENTIALS = 'creds-prod'
MY_ENVIRONMENT = 'staging'
break
default:
println("Branch value error: " + branch)
currentBuild.getRawBuild().getExecutor().interrupt(Result.FAILURE)
}
}
}
}
stage("Publish") {
steps {
withCredentials([[$class: 'AmazonWebServicesCredentialsBinding', credentialsId: env.MY_CREDENTIALS, accessKeyVariable: 'AWS_ACCESS_KEY_ID', secretKeyVariable: 'AWS_SECRET_ACCESS_KEY']]) {
sh 'make takecareofit'
}
}
}
}
}