How can I set a schema without having to run the preview?

Data Collector performs a schema-on-read, so you should preview the data at least once to capture the schema. There is an option in the Preview Configuration to save the preview record schema which Determines whether the preview record schema is saved in the pipeline configuration. By default, the schema is not saved. When you close and then reopen the pipeline, the Schema tab for each pipeline stage is empty. You must run preview again to update the input and output schema. When selected, the schema is saved after the pipeline is closed.