Why YouTube tutorials mostly shows postgres and not MySQL/mariaDB?

Hypster recommendations

Because unless your doing some overly complicated or trying to make a monolithic query continue to run when it starts to fail, MySQL is straight forward enough.

1. No need to worry about the storage engine
2. Utf-8 means utf-8
3. There is an dispute of the floss community with oracle about mysql.
4. You have builtin support for many datatypes:
- uuid
- serial
- time (date, time datetime with and without timezone)
- ipaddresses
- json and xml
- geometric types
- an extension provides si and non-si units as types.

Both are complex things.

All have a very great and detailed documentation that is far more detailed and correct than most stuff on stackofshit or YouTube.

@stop That's wrong. Some of the stuff exists in MySQL / MariaDB, and Postgres is on it's way to pluggable storage

;)

All in all, both have nowadays a good feature parity that is more than sufficient for anything OOTB. Configuration needs to be done properly, security too.

If you _really_ want to use a database _fully_… you won't look on shit overflow or YouTube. And most likely you will not use an ORM or stuff like that...

@stop they both have json but Postgres has jsonb with gin indexes, which is nice and very fast compared to MySQLs implementation.

People always describe sql harder than it is, like there is some magic to it; when in reality it’s the easiest of the coding languages…you only have 4 commands, the rest is how and where.

@atrabilious i never got my head to understand joins enough so i can use them correctly. I always end up with strange warnings, wrong results or duplicate results.

@atrabilious technically you could do the same with an full text index afaik in MySQL :-P as GIN is just an inverted index.

And SQL is just one part of the cake, imho. SQL itself is easy, getting the database to run the SQL how you like it is hard.

@stop

When you don't understand joins... Then SQL is hard. As joins are one of the integral parts of SQL.

@stop left - 1 to many (for 1 row in first table bring me every matching row in 2nd table), inner - 1 to 1 (only return when you find a match in both) right - many to one, (for every row in first table go find one match in the 2nd). The complication comes on join 6 and is it a left join after a right and two inners? 😝

It’s not all simple, a lot of people can’t wrap their head around recursion either, but for the most part it’s the easiest thing I work with.

@IntrusionCM with jsonb and containment you don’t even need the gin index, but it makes not having to put the whole path in super easy and still fast… and fulltext isn’t nearly as efficient, but granted that works. I use MySQL and Postgres, for different projects- they both have their uses, but I always end up with faster response times with docs in pg. I keep getting pushed to use mongo atlas, and yet I resist it.

@atrabilious Yes, Full text search is implemented poorly in MySQL and not based on an inverted index afaik. That you'll get faster speed with Postgres isn't a miracle, either.

The parallelism introduced in Postgres in recent versions is indeed a big plus in "raw" performance vs MySQL. As long as you have enough threads / ram as resources, postgres parallelism does wonders.

https://dev.mysql.com/doc/refman/...

Okay, they're based on an inverted index.

My bad.