I work at a large enterprise. At one point the large enterprise decided to invest heavily in an ahem non-relational model of storing data. This has proven to

I work at a large enterprise. At one point the large enterprise decided to invest heavily in an ahem non-relational model of storing data. This has proven to be pretty dumb. It's both less performant and less reliable than the relational databases it purports to replace. On top of this, any interesting application of the data besides basic CRUD is either slow and a total pain in the ass or just not possible. SQL isn't going away because it's valuable to exploit relationships between data and because it's pretty good at doing that. Anything that tries to replace it, eventually looks suspiciously like it. (I'm looking at you, MongoDB aggregate functions)