A lot of goodies around JSONB additions. In 9.4, there wasn't much support for updates to JSONB values, but now there will be some support for document-esque operations:
- jsonb || jsonb (concatenate / overwrite) : combine 2 jsonb objects and if there's overlap, values are replaced at the highest level
- jsonb - text / int (remove key / array element) : remove key or an array element
- jsonb #- text / int (remove key / array element in path) : remove key/array at path
- jsonb_set function : update value at path
I'm a bit sad that at my current job our problem set is much better suited for a distributed key-value store; I miss postgres, but it really is the wrong tool for this particular problem.
Edit: I'm thinking of more DBA type work and assisting with/contributing to drivers/wrappers.
Think of any "big data" company -- but moreso - think of anyone who is using Amazon Redshift for BI...
But really any company who can say they have "a data pipeline" will be a good candidate.
If you have very deep Postgres knowledge, then make sure you have strong ability to ferry data in and out of the DB esp. on AWS.
There are so many people who are looking to move data into and out of Redshift, RDS or other standalone systems. Any company who uses the term ETL will
Also - there will be a good number of people who want to migrate, at least in testing, from Postgres to Aurora (mysql based - but instance-less) -- so this will be a good skill.
EDIT: I'm currently looking.
"There’s a few updates to JSONB. The one I’m most excited about is making JSONB output in psql read much more legibly."
This is good, but there's a link to other much more useful JSONB functions: https://wiki.postgresql.org/wiki/What's_new_in_PostgreSQL_9....