### Terraform Version ```shell Terraform v1.13.3 on windows_amd64 + provider registry.terraform.io/hashicorp/google v6.49.2 ``` ### Terraform Configuration Files The check on for_each is need when there is no entries on yaml file. The problem is that this check produces the error ahead when I have two cloud runs with different specs. ```terraform locals { runjob = yamldecode(file("./yaml/runjob.yaml")) } module "runjob" { source = "git::https://github.com/danilomnds/terraform-gcp-cloud-run-job?ref=v1.1.0" # check for_each = (local.runjob.runjob) == null ? {} : local.runjob.runjob name = lookup(each.value, "name", null) == null ? each.key : lookup(each.value, "name", null) template = each.value.template location = each.value.location labels = each.value.labels annotations = lookup(each.value, "annotations", null) client = lookup(each.value, "client", null) client_version = lookup(each.value, "client_version", null) launch_stage = lookup(each.value, "launch_stage", null) binary_authorization = lookup(each.value, "binary_authorization", null) start_execution_token = lookup(each.value, "start_execution_token", null) run_execution_token = lookup(each.value, "run_execution_token", null) project = var.project members = lookup(each.value, "run_execution_token", []) scheduler_jobs_admin = lookup(each.value, "scheduler_jobs_admin", false) deletion_protection = lookup(each.value, "deletion_protection", null) } output "runjob" { value = { for name, id in module.runjob : name => id.id } } ...terraform config... ``` Trying to create two cloud runs. One does integrate with a VPC while the other one no. ```yaml runjob: runjob-bigdata-dev-1: name: "runjob-bigdata-dev-1" location: "us-east1" template: template: containers: name: "democontainer" image: "us-docker.pkg.dev/cloudrun/container/job:latest" resources: limits: cpu: 1000m memory: "1Gi" vpc_access: network_interfaces: network: "tim-cloud-hdev" subnetwork: "us-east1" labels: system: "your system" runjob-bigdata-dev-2: name: "runjob-bigdata-dev-2" location: "us-east1" template: template: containers: name: "democontainer" image: "us-docker.pkg.dev/cloudrun/container/job:latest" resources: limits: cpu: 1000m memory: "1Gi" labels: system: "your system" ``` ### Debug Output ``` 2025-10-02T12:38:03.396-0300 [TRACE] dag/walk: upstream of "module.runjob.output.id (expand)" errored, so skipping 2025-10-02T12:38:03.396-0300 [TRACE] dag/walk: upstream of "module.runjob.google_project_iam_member.CustomCloudRunDeveloper" errored, so skipping 2025-10-02T12:38:03.396-0300 [TRACE] dag/walk: upstream of "module.runjob.google_project_iam_member.CloudSchedulerAdmin" errored, so skipping 2025-10-02T12:38:03.396-0300 [TRACE] dag/walk: upstream of "provider[\"registry.terraform.io/hashicorp/google\"] (close)" errored, so skipping 2025-10-02T12:38:03.396-0300 [TRACE] dag/walk: upstream of "module.runjob (close)" errored, so skipping 2025-10-02T12:38:03.396-0300 [TRACE] dag/walk: upstream of "output.runjob (expand)" errored, so skipping 2025-10-02T12:38:03.396-0300 [TRACE] dag/walk: upstream of "root" errored, so skipping Releasing state lock. This may take a few moments... ╷ │ Error: Inconsistent conditional result types │ │ on runjob.tf line 7, in module "runjob": │ 7: for_each = (local.runjob.runjob) == null ? {} : local.runjob.runjob │ ├──────────────── │ │ local.runjob.runjob is object with 2 attributes │ │ The true and false result expressions must have consistent types. The 'false' value includes object attribute "runjob-bigdata-dev-1", which is absent in the 'true' value. ╵ 2025-10-02T12:38:04.182-0300 [DEBUG] provider.stdio: received EOF, stopping recv loop: err="rpc error: code = Unavailable desc = error reading from server: EOF" 2025-10-02T12:38:04.192-0300 [INFO] provider: plugin process exited: plugin=.terraform/providers/registry.terraform.io/hashicorp/google/6.49.2/windows_amd64/terraform-provider-google_v6.49.2_x5.exe id=31192 2025-10-02T12:38:04.192-0300 [DEBUG] provider: plugin exited [terraform.log](https://github.com/user-attachments/files/22663276/terraform.log) ``` ### Expected Behavior I want to check whether the runjob.yaml file is empty. If it's empty, the system should either: - Do nothing, if no resources were previously created. - Delete the last remaining resource, if this is the final cleanup step. if yaml is like this ```yaml runjob: ``` On terraform "do nothing" ```terraform for_each = (local.runjob.runjob) == null ? {} : local.runjob.runjob ``` If yaml is like this ```yaml runjob: runjob-bigdata-dev-1: name: "runjob-bigdata-dev-1" runjob-bigdata-dev-2: name: "runjob-bigdata-dev-2" ``` On terraform "go ahead" on resource creation/deletion/update ```terraform for_each = (local.runjob.runjob) == null ? {} : local.runjob.runjob ``` ### Actual Behavior ```PS PS C:\Users\F8082011\OneDrive - TIM\TIM\Cloud\Azure\Azure Devops\gcp-fasttrack-bigdata\environments\fqa\runjob> terraform plan ╷ │ Error: Inconsistent conditional result types │ │ on runjob.tf line 7, in module "runjob": │ 7: for_each = (local.runjob.runjob) == null ? {} : local.runjob.runjob │ ├──────────────── │ │ local.runjob.runjob is object with 2 attributes │ │ The true and false result expressions must have consistent types. The 'false' value includes object attribute "runjob-bigdata-dev-1", which is absent in the 'true' value. ``` ### Steps to Reproduce 1. terraform init 2. terraform apply ### Additional Context _No response_ ### References _No response_ ### Generative AI / LLM assisted development? _No response_