The FileBatchSource currently always set the default schema as the output schema. This is not always true since we can add a path field to the output schema. If the following plugin tries to get the input schema, this path field will be missing.
To reproduce, use a file source and a run transform, specify path field in the file source and add it to the fields to process in the run transform. The pipeline will not get deployed and a RuntimeExeption is always thrown, saying this path field is missing from input schema.