I quickly found dealing with time series reports dramatically easier. There's the possibility I was just doing it wrong with MySQL. It seemed so much more expressive for things like writing a query based on an interval and perhaps filling empty cells with default values for example. I also found it very intuitive to connect a UI to our API which would convert various parameters to dynamic queries. That was a game changer; almost all reports could be run by anyone in the company without asking someone to write a query for them. The result was remarkably reliable and easy to maintain. the MySQL version I refactored was much harder to deal with, slower, and only getting worse despite our best efforts.
There were very cool things I couldn't find a way to do in MySQL which eventually made the switch an amazing change. It was trivial to set up aggregated time series tables, and very easy to roll that data up in order to speed up queries with a coarser granularity. At the time I recall window functions making this really clean and easy. It was possible in MySQL (and maybe trivial today, too) but it felt like a massive kludge every step of the way.
This is actually what lead us to consulting someone who could implement the best solutions to these issues that they knew. They encouraged us to adopt Postgres before doing any work, we didn't listen, then as they wrapped up the project they once again advised we leave MySQL behind. He then sent some examples of how we might make the migration and why it would be better. Months later we had made the migration entirely and were very, very glad.
That guy was an AWESOME DBA, and he changed the way I value someone who is good with data and databases. He transformed our team's ability to grow our product, dramatically improved our ability to deliver, and instilled a ton of knowledge in a short time. I can only dream of imparting that to other teams!
Overall I'd say that Postgres provided tools that made working with data easier. I'm not sure if there are any specific features besides a cleaner, simpler,
Weird that that got cut off. I think I intended to say "a cleaner, simpler, more efficient set of tools for working with the type of data I had". And of course, right tool for the right job and all, perhaps there are plenty of cases where MySQL is better. I just haven't found them yet, and probably won't any time soon.
There were very cool things I couldn't find a way to do in MySQL which eventually made the switch an amazing change. It was trivial to set up aggregated time series tables, and very easy to roll that data up in order to speed up queries with a coarser granularity. At the time I recall window functions making this really clean and easy. It was possible in MySQL (and maybe trivial today, too) but it felt like a massive kludge every step of the way.
This is actually what lead us to consulting someone who could implement the best solutions to these issues that they knew. They encouraged us to adopt Postgres before doing any work, we didn't listen, then as they wrapped up the project they once again advised we leave MySQL behind. He then sent some examples of how we might make the migration and why it would be better. Months later we had made the migration entirely and were very, very glad.
That guy was an AWESOME DBA, and he changed the way I value someone who is good with data and databases. He transformed our team's ability to grow our product, dramatically improved our ability to deliver, and instilled a ton of knowledge in a short time. I can only dream of imparting that to other teams!
Overall I'd say that Postgres provided tools that made working with data easier. I'm not sure if there are any specific features besides a cleaner, simpler,