Note the following passage:

"It's clear to me that a certain proportion of application developers are allergic to putting business logic in the database system, despite the fact that batch processing logic is often a great deal faster when implemented that way. Perhaps the fact that clients are now able to cede full control of transaction scope to procedural code running on the database server (code that can be written in a familiar scripting language) will help application developers to get over the understandable aversion. Hope springs eternal."

Now, if I could only have a pljulia... so that I could run statistical analysis (packaged as a Julia function) from a query. With PostgreSQL the request queue is never ending ;)

Curious to see if the emphasis on parallelism will affect Citus Data's business model?

A bit off topic but I wish postgres came with a good desktop app. pgadmin 4 is just abysmal compared to e.g. MySQL Workbench.

I'm fairly newb to PostgreSQL, and am not an expert (just a user) in many SQL systems. But if Postgres' typecasting methodology is unique to it, by jove it's a great point.

What other features should I learn about (comparisons to MySQL/Oracle)? Any reason to focus on Postgres/Greenplum over a Hadoop hive/impala setup?

Anyone know what the filter evaluation redesign they did in version 10 was?

> The v10 improvement essentially flattened the representation of expression state and changed how expression evaluation was performed.

It sounds like they were using an expression tree and changed it to something like truth tables, which are essentially "flat" expression representations that can support any boolean expression.

i am enjoying it much more than sql server

How is it in regards to Vacuuming and DB size?

I'm surprised covering indices is only now appearing in 11. It's fairly commonplace in other databases.

Would be nice to see some JIT benchmarks with sample queries and the JIT speedup.

Love Pirsig's analogies and can't resist to point out that a "classic approach" will always beat an ignorant hype-driven bullshit in the long run.

Informix (now dead?) vs. Oracle is another classic example from the past.

Oracles's domination absolutely does not imply its technical superiority, actually popularity is usually a bad thing (hello, Ayn Rand) - Java and junk food are insanely popular.