Last year, I attended an event during which SQL trainer and consultant Markus Winand gave an aptly titled talk on Modern SQL: Evolution of a dinosaur (the linked video is not from that event but it shows the same talk given at another event).
Markus runs a website called Modern SQL, which provides an impressive range of SQL tips, best practices, descriptions of both highly useful and lesser known SQL features and comparisons between the various relational database management systems (RDBMS) available on the market.
I consider myself reasonably well-versed in SQL and its somewhat lesser known features such as window functions, common table expressions (CTE) and recursion but from both that talk and Modern SQL I’ve learned about quite a few new approaches and techniques such as the very promising row pattern matching feature introduced with SQL:2016.
Basically, this feature allows you to define patterns and use those to match rows with regular expressions! This should prove eminently useful when working with logs or similar time series data and reactive application patterns.