The spark.read.json internally uses TextInputFormat, which processes one line at a time. Therefore, one JSON record cannot be on multiple lines. It would be a valid JSON format if you use multiple lines, but it will not work with Spark and will throw an exception.

It is allowed to have more than one object in a line. For example, you can have the information of two persons in one line as an array, as follows:

[{"firstName":"Barack", "lastName":"Obama"},{"firstName":"Bill", "lastName":"Clinton"}]

This recipe concludes the saving and loading of data in the JSON format using Spark.