I was hoping this was an implementation of the schemaless indexing [1], which is the foundation for Azure DocumentDB.
That design allows arbitrary nested JSON data to be indexed using inverted indexes on top a variation of B-trees called Bw-trees, and seems like a nice way of indexing data automatically in a way that preserves the ability to do both exact and range matching on arbitrarily nested values.
That design allows arbitrary nested JSON data to be indexed using inverted indexes on top a variation of B-trees called Bw-trees, and seems like a nice way of indexing data automatically in a way that preserves the ability to do both exact and range matching on arbitrarily nested values.
[1] https://www.vldb.org/pvldb/vol8/p1668-shukla.pdf