3 ms·

> "I have no idea if he is aware of other types of joins despite being the best database/SQL architect I’ve ever met or heard of." That just can't be right. Gr

by smt88 5y ago

> "I have no idea if he is aware of other types of joins despite being the best database/SQL architect I’ve ever met or heard of." That just can't be right. Great SQL people understand set theory, query planning, and a lot of arcane RDBMS internals that most of us don't have to get into. I studied databases in college and have been writing SQL for 25 years, and I am regularly blown away by the knowledge people demonstrate on Stack Overflow. There is just no way a great DB architect or SQL programmer would only use left joins. The person who wrote this is likely just early enough in their career that they don't know what a truly great DB expert looks like.

bryanrasmussen 5y ago
>The person who wrote this is likely just early enough in their career I suspect that they were 18 in 2012 based on profile, so their career has been 10 years long so far. As far as no great would do X, not sure (as I avoid the relational DB side of things), but sometimes experts in a thing decide to only use a subset of the capabilities for personal preferences or because they feel it makes them more productive. So maybe in this guy's mind there was an SQL: the Good parts! which only included left joins.
- smt88 5y ago
  I think you could easily get 10 years into a web dev career and never meet a truly fantastic DB person. Most companies just aren't successful enough to need one. I am a reformed MySQL user, but one of my overly-normalized MySQL DBs has been running in production for almost 10 years without any optimization of queries, tables, or indices (beyond some naive indices I created when deploying it). My point is that you can throw an excellent RDBMS onto some decent hardware and serve millions of requests for a CRUD web app without really thinking too hard about it.
athrowaway3z 5y ago
>That just can't be right. Great SQL people understand set theory, query planning, and a lot of arcane RDBMS internals that most of us don't have to get into. Your likely misinterpreting what the author means with great. If your not at a scale where a team is pushing the micro optimization boundary, there is tremendous value in the ability to keep everything simple. Having the entire org operate using the same frame is great. What's more, it's not uncommon for great simplicity to be mistaken for 'obvious', and excessive complexity be mistaken for great work.
- BiteCode_dev 5y ago
  Which is also why ORM are so popular despite being suboptimum. Not because they save you from writting SQL, but because they force people to standardize : to define a single point of entry for they data, to centralize the schema, integraty checks or signals, and of course, normalize the way the query are written. This has greater value than perf for a lot of orgs.
jawns 5y ago
If this architect was that wedded to LEFT only, one would think he would have been outspoken about why. Most of the great programmers I've known have had two characteristics: They'll tell you very confidently about what you SHOULD do, and they'll tell you very loudly about what you SHOULD NOT do.
- kcplate 5y ago
  I have always said that if if you put 10 great programmers into a room together you would have 10 people who are absolutely convinced they are the smartest person in the world surrounded by the 9 dumbest people in the world.

by smt88 5y ago

bryanrasmussen 5y ago

>The person who wrote this is likely just early enough in their career I suspect that they were 18 in 2012 based on profile, so their career has been 10 years long so far. As far as no great would do X, not sure (as I avoid the relational DB side of things), but sometimes experts in a thing decide to only use a subset of the capabilities for personal preferences or because they feel it makes them more productive. So maybe in this guy's mind there was an SQL: the Good parts! which only included left joins.

smt88 5y ago
I think you could easily get 10 years into a web dev career and never meet a truly fantastic DB person. Most companies just aren't successful enough to need one. I am a reformed MySQL user, but one of my overly-normalized MySQL DBs has been running in production for almost 10 years without any optimization of queries, tables, or indices (beyond some naive indices I created when deploying it). My point is that you can throw an excellent RDBMS onto some decent hardware and serve millions of requests for a CRUD web app without really thinking too hard about it.

athrowaway3z 5y ago

>That just can't be right. Great SQL people understand set theory, query planning, and a lot of arcane RDBMS internals that most of us don't have to get into. Your likely misinterpreting what the author means with great. If your not at a scale where a team is pushing the micro optimization boundary, there is tremendous value in the ability to keep everything simple. Having the entire org operate using the same frame is great. What's more, it's not uncommon for great simplicity to be mistaken for 'obvious', and excessive complexity be mistaken for great work.

BiteCode_dev 5y ago
Which is also why ORM are so popular despite being suboptimum. Not because they save you from writting SQL, but because they force people to standardize : to define a single point of entry for they data, to centralize the schema, integraty checks or signals, and of course, normalize the way the query are written. This has greater value than perf for a lot of orgs.

jawns 5y ago

If this architect was that wedded to LEFT only, one would think he would have been outspoken about why. Most of the great programmers I've known have had two characteristics: They'll tell you very confidently about what you SHOULD do, and they'll tell you very loudly about what you SHOULD NOT do.

kcplate 5y ago
I have always said that if if you put 10 great programmers into a room together you would have 10 people who are absolutely convinced they are the smartest person in the world surrounded by the 9 dumbest people in the world.