So here's the deal... The best way to create and debug a U-SQL script is to run it directly from a Data Lake Analytics job. We suggest you read/implement what's written in the next section (run U-SQL from a job in the Data Lake Analytics) before you continue here. In the next section, you will find all the configurations that are required to run U-SQL as well as the U-SQL script itself.

We will use the same pipeline we have previously created to add a U-SQL task for data manipulation in the data lake. Reopen the pipeline and open the Data Lake Analytics activity. Then drag and drop the U-SQL task:

You can modify the name of the U-SQL:

Next, in the U-SQL task, click on the ADLA Account tab and click on New:

A new window will open; you may change the name of the new ADLA linked service and choose the new Data Lake Analytics resource.

Scroll down. The tenant ID will be automatically populated. The Service Principal ID needs to be filled in. Details are at https://docs.microsoft.com/en-us/azure/data-factory/transform-data-using-data-lake-analytics#azure-data-lake-analytics-linked-service.