There are 3 official built-in extractors that allows you to analyze data contained in CSV, TSV or Text files.
But MSFT also released some additional sample extractors on their Azure GitHub repo that deal with Xml, Json and Avro files. I have used the Json extractor in production as it is really stable and useful.
The JSON Extractor treats the entire input file as a single JSON document. If you have a JSON document per line, see the next section. The columns that you try to extract will be extracted from the document. In this case, I'm extracting out the _id and Revision properties. Note, it's possible that one of these is a further nested object, in which case you can use the JSON UDF's for subsequent processing.
REFERENCE ASSEMBLY [Newtonsoft.Json];
REFERENCE ASSEMBLY [Microsoft.Analytics.Samples.Formats];
//Define schema of file, must map all columns
@myRecords =
EXTRACT
_id string,
Revision string
FROM @"sampledata/json/{*}.json"
USING new Microsoft.Analytics.Samples.Formats.Json.JsonExtractor();