I would like to be able to get custom output from an "Execute Pipeline Activity". During the execution of the invoked pipeline, I capture some information in a variable using the "Set Variable" activity. I would like to be able to use that value in the master pipeline.
I know that the master pipeline can read the invoked pipeline's name and runId using "@activity('InvokedPipeline').output," but those are the only properties available.
I have the invokable pipeline because it's configurable to be used by multiple other pipelines, assuming we can get the output from it. It currently consists of 8 activities; I would hate to have to duplicate them all across multiple pipelines just because we can't get the output from an invoked pipeline.
Reference: Execute Pipeline Activity
[
{
"name": "MasterPipeline",
"type": "Microsoft.DataFactory/factories/pipelines"
"properties": {
"description": "Uses the results of the invoked pipeline to do some further processing",
"activities": [
{
"name": "ExecuteChildPipeline",
"description": "Executes the child pipeline to get some value.",
"type": "ExecutePipeline",
"dependsOn": [],
"userProperties": [],
"typeProperties": {
"pipeline": {
"referenceName": "InvokedPipeline",
"type": "PipelineReference"
},
"waitOnCompletion": true
}
},
{
"name": "UseVariableFromInvokedPipeline",
"description": "Uses the variable returned from the invoked pipeline.",
"type": "Copy",
"dependsOn": [
{
"activity": "ExecuteChildPipeline",
"dependencyConditions": [
"Succeeded"
]
}
]
}
],
"parameters": {},
"variables": {}
}
},
{
"name": "InvokedPipeline",
"type": "Microsoft.DataFactory/factories/pipelines"
"properties": {
"description": "The child pipeline that makes some HTTP calls, gets some metadata, and sets a variable.",
"activities": [
{
"name": "SetMyVariable",
"description": "Sets a variable after some processing from other activities.",
"type": "SetVariable",
"dependsOn": [
{
"activity": "ProcessingActivity",
"dependencyConditions": [
"Succeeded"
]
}
],
"userProperties": [],
"typeProperties": {
"variableName": "MyVariable",
"value": {
"value": "@activity('ProcessingActivity').output",
"type": "Expression"
}
}
}
],
"parameters": {},
"variables": {
"MyVariable": {
"type": "String"
}
}
}
}
]
Yes. An activity output can be consumed in a subsequent activity with the @activity construct.
unfortunately we don't have a way to pass the output of an activity across pipelines. Right now pipelines don't have outputs (only activities).
An input dataset represents the input for an activity in the pipeline, and an output dataset represents the output for the activity. Datasets identify data within different data stores, such as tables, files, folders, and documents. After you create a dataset, you can use it with activities in a pipeline.
Hello Heather and thank you for your inquiry. Custom outputs are not an inbuilt feature at this time. You can request/upvote for the feature in the Azure feedback forum. For now, I do have two work-arounds.
Utilizing the invoked pipeline's runID, we can query the REST API (using Web Activity) for the activity run logs, and from there, the activity outputs. However before making the query, it is necessary to authenticate.
REST call to get the activities of a pipeline
For authentication I reccomend using the Web Activity to get an oauth2 token. The URL would be https://login.microsoftonline.com/tenantid/oauth2/token
. Headers "Content-Type": "application/x-www-form-urlencoded"
and body "grant_type=client_credentials&client_id=xxxx&client_secret=xxxx&resource=https://management.azure.com/"
. Since this request is to get credentials, the Authentication setting for this request is type 'None'. These credentials correspond to an App you create via Azure Active Directory>App Registrations. Do not forget to assign the app RBAC in Data FActory Access Control (IAM).
Another work-around, has the child pipeline write its output. It can write to a database table, or it can write to a blob (I passed the Data Factory variable to a Logic App which wrote to blob storage), or to something else of your choice. Since you are planning to use the child pipeline for many different parent pipelines, I would reccomend passing the child pipeline a parameter which it uses to identify the output to the parent. That could mean a blob name, or writing the parent runID to a SQL table. This way the parent pipeline knows where to look to get the output.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With