Hacker News Reader

Comment by bmpafa | Hacker News Reader

These are great. I feel like no matter how many PG tips & tricks I learn, there's always something new out there.

I recently started learning PostgreSQL after ~2 yrs of relying on BaaS etc. (mostly of the NoSQL variety).

I have to say: I regret not learning it earlier. Stuff like constraints, triggers, views, etc., are game-changers, but even just being able to write queries instead of navigating a proprietary GUI was worth the few weeks it took to learn.

It was also a boon for data analysis work. Whereas before I'd take a few hours cleaning some data with JavaScript or Python, now I just write a SQL query or two. It's faster, it feels more powerful, and it's _substantially_ less fragile than one-off data cleaning scripts.

Anyway: for those that started out like I did, afraid of SQL, dive on in.

View on HN

How did you learn? Old-fashioned trial and error + googling?

Read through the docs

The PG docs are surprisingly good. Good enough to learn from. I remember learning my first SQL back in like 1999 or 2000 by reading the PG docs in the back of my parents minivan. Good times!

I don't mean to knock the pg docs but I honestly come away reading them with the feeling "what did I just read?". They seem dense and dry. So personally I appreciate the articles like this to help me really learn. I will skip over official documentation and go to this sort of content.

> They seem dense and dry.

I consider that a feature. I loathe wasting my time by having to dig through layers of prose to get some fact.

But they are probably not suitable as a tutorial, they only make sense when you already know relational database basics and need info on pg-specifics.

It's a personal preference I guess. I'd rather hit up stack overflow and find a common English answer to a similar problem that usually includes an example. Django suffers from the same thing... excellent docs, that just seem unapproachable day to day.

Basically when I Google something, I'm skipping the first 2-3 official links that in fact contains the info I need, but instead clicking the one that summarized it more plainly

If it doesn't actually resonate with you, and you don't come away with the understanding you need, it's most certainly not a 'feature' but a 'bug'. I think some projects see dense/thicks docs that keep a high barrier to entry as a feature itself as well.

If they're not approachable, they're not 'excellent'.

I've been a django developer for a decade and keep hearing from others how great their docs are. I barely ever use them. My biggest complaint is they give examples without any context. I can't look up a feature without knowing 12 other things before

Yup-3 parts elbow-grease, 2 parts tears & sweat. I had a lot of data to work through, so I used that as an excuse. Not having to worry about performance, schemes, servers, users, etc., made learning a lot easier, I think.

I started by reading the tutorial in the PG docs[1]. It was _very_ well-written & accessible--def. recommend.

Later, I picked-up the book "Mastering PostgreSQL in Application Development"[2]. It was pretty solid, and seemed one of the only books aimed at developers (as opposed to DBAs).

[1] https://www.postgresql.org/docs/10/static/tutorial.html

[2] https://masteringpostgresql.com/

After mostly relying on ORMs for years, I wanted to brush up on SQL, so I wrote this...

https://github.com/amorphid/learn_sql_ruby

Basically I just used the pg gem for Ruby to execute a bunch of raw SQL commands and wrote tests to check to see if the output was correct. It taught me a bunch in just a few days!

ORM is much easier when you know what is going on and can dove into a query plan. I have noticed most devs that struggle with ORM and or performance tend to just not understand basic RDBMS/SQL stuff. Indexes, joins, how to tell ORM to prefetch, reviewing query plans when the performance matters.

I also recently began learning SQL properly, and I also regret not having done so earlier.

The PG docs/tutorial is really great to get started.

After I worked through those, I feel it was essential to work my way through pgexercises.com. It's really easy to think you get it, but actually having to puzzle through stuff is obviously quite different.

When I reached some of the more complicated subquery stuff, I decided to get 'Learn SQL' by Alan Beaulieu (recommended by pgexercises), and I just worked my way from the start all over again. Book's excellent too.

I do find that I hit a 'barrier' when all the basic parts are introduced and it becomes all about translating a goal to 'thinking in SQL', especially when there seem to be multiple valid approaches. But the combination of reading a ton, then doing the exercises, then re-reading and trying to play around with the Learning SQL db (Sequel Pro is great for Mac, as is Postico (Postgres)) has been very effective so far.

I found this recent book by one of the Postgres contributors helpful: https://masteringpostgresql.com/

Reply on news.ycombinator.com