Do you prefer Postgres over MySQL? Why?

I use Postgres because of jsonb, postgis, recursive CTEs, the great functions that come with it etc. I realise MySQL has equivalents now but they're not so better I'd change back.

TRUNCATE TABLE foo RESTART IDENTITY should fix what makes you crazy :)

I've used both, but to be honest I usually stick with MySQL.
It seemed like the obvious choice several years ago when I started and I never questioned my choice since then. I've yet to find a use case that would make me switch. And I guess for my general websites there's hardly a reason to make the switch.

I've only done a couple of projects with them, but moved to Postgres because MySql now has mixed licensing that include GNU and proprietary agreements.

I have used both for websites with relatively high traffic.

In my experience, MySQL was crashing and using up server's RAM and CPU much more than PostgreSQL. PostgreSQL seemed to be more stable and better performing. In addition, PostgreSQL has JSON fields, PostGIS extensions for mapping, UUID fields, and some other cool features.

The only thing that seems to me better about MySQL is that there are more client apps to manage it and usually they are more intuitive. Some of them worth mentioning are Sequel Pro (https://www.sequelpro.com/) and HeidiSQL (https://www.heidisql.com/).

read somewhere on Twitter. (couldn't find the tweet)

Postgres is the Linux of databases.

Postgres all the way. The array and JSON fields are great, but my favorite feature is that DDL scripts are transactional.

I like Postgres for its Array and JSON fields. I do sometimes need these in my apps, but I'm quite cautious about NoSQL databases.

Most features that distinguish database engines are very rarely visible for simple applications. Therefore you can pick whatever SQL or NoSQL database you prefer and that is fine. However, if you venture into the territories of 100+ billion-row tables, 100+ TB databases, 2000+ line complex SQL queries, 200k+ active connections then you start to reevaluate what is important. In my line of work, the question always was whether we use Oracle and pay the price or experiment with Postgresql. Over the years, Postgres was more and more a natural choice for these large analytical applications. Few features which helped us are:

• extensibility with various languages: I worked with an application where half of the application logic implemented in stored function in 3 different languages.
• PostGIS, parallel queries, materialized views, database clusters, query on deeply nested JSON structures, sharding, publish/notify event system, backgound processes
• monitoring and observability tools,
• fine tuning of query planner and performance (I regularly beat NoSql engines performance wise)
  Also Postgres extensibility helps to create your own data or index types which led to derivative products recently like: CitusDB, Timescale DB that I used. I think about PG more like an application development platform than a database.

I used to use MySQL but I didn't see much difference between MySQL and PostgreSQL. I made the switch to NoSQL and I use Mongo now because I like the document-oriented storage.

I used to stand behind MySQL until I had to deploy it a couple times. I dunno how much blame I can put on MySQL but after spending several nights failing to get external connections to work I ditched in for Postgres. Postgres was a lot easier to set up and I haven't looked back since.

Practical reason: several managed DB services only offer Postgres (eg Heroku, Digital Ocean, etc). This is not true across the board as many also offer MySQL.

Personally: we used Postgres in my Databases course in college and I've stuck with it since.

No preference here, Postgres is doing a lot of great things. MySQL / MariaDB has served me well in the past as well.

Regarding your testing issue, perhaps delete the data outright instead of truncating the table?

It's because all the developers are using it. All the well-defined docs like DigitalOcean and other popular docs promoting PostgreSQL.

It's not that hard to set up and readily available in the Linux machine. In case if you want some No SQL functionality you can use PostgreSQL as no-sql.