|
here’s a go tutorial for you: go get a new job
|
# ? Dec 3, 2021 13:02 |
|
|
# ? May 9, 2024 03:06 |
insane snipe energy
|
|
# ? Dec 3, 2021 13:11 |
|
cinci zoo sniper posted:insane snipe energy
|
# ? Dec 3, 2021 13:58 |
|
cinci zoo sniper posted:insane snipe energy
|
# ? Dec 3, 2021 14:44 |
|
Private Speech posted:I actually find the first easier to read than the second what the heck, for_each is the one not to like
|
# ? Dec 3, 2021 15:21 |
|
cinci zoo sniper posted:insane snipe energy
|
# ? Dec 3, 2021 15:28 |
|
jesus WEP posted:here’s a go tutorial for you: go get a new job new title?
|
# ? Dec 3, 2021 15:40 |
if it fits im all for it
|
|
# ? Dec 3, 2021 15:42 |
|
cinci zoo sniper posted:insane snipe energy
|
# ? Dec 3, 2021 16:02 |
|
cinci zoo sniper posted:insane snipe energy
|
# ? Dec 3, 2021 16:30 |
|
NihilCredo posted:sql is a bit of a tragedy because the "runtimes" are so freaking good but it's impossible to fix the relatively minor design flaws in its frontend for the most part i think sql is fine and i wouldnt really change much. nulls are a sign of bad data or bad data design which cant be fixed by language. on the other hand, ansi nulls are objectively wrong and microsoft should have punted on implementing them at all in sql server.
|
# ? Dec 3, 2021 16:32 |
|
Xarn posted:Apparently I get to maintain a bunch of Go code, because the teams using our library from Go cannot be trusted to do the sane thing and write common utility functions once, so there are now 4 subtly different versions of conceptually same code between different go projects. besides the snipe, the "tour of go" on the official golang web site is actually a great introduction
|
# ? Dec 3, 2021 16:39 |
|
Arcsech posted:these functions that are semantically clearer than looping
|
# ? Dec 3, 2021 16:53 |
|
Sagacity posted:golang user spotted
|
# ? Dec 3, 2021 16:54 |
|
Shaggar posted:for the most part i think sql is fine and i wouldnt really change much. nulls are a sign of bad data or bad data design which cant be fixed by language. both your complains derive from the same flaw, the conflation of 'undefined' nulls (from missing outer joins) and 'null' nulls (to represent nonexisting values, which is a legitimate data design, it just shouldn't be the default) ansi nulls are correct for the former, because you don't want 'a.foo = b.foo' to return true when there is no match in either a or b imagine if LEFT JOIN returned 'undefined' instead of 'null' when the condition failed, and the special rules only applied to undefined. now if you filter for 'a.foo = b.foo' you get: - if there's no record in either a or b => false - if there's a record in both a and b, but both columns are intentionally null => true which is what you would expect
|
# ? Dec 3, 2021 17:32 |
|
i dont like that you have to use is for null, its annoying
|
# ? Dec 3, 2021 17:43 |
|
oh advent of code doesn't show anywhere how many attempts it's taken you to solve the problem? this is bullshit. how am i supposed to brag??
|
# ? Dec 3, 2021 17:49 |
|
PIZZA.BAT posted:oh advent of code doesn't show anywhere how many attempts it's taken you to solve the problem? this is bullshit. how am i supposed to brag?? mystes fucked around with this message at 17:58 on Dec 3, 2021 |
# ? Dec 3, 2021 17:54 |
|
mystes posted:You probably realize this but if you want to go down the dumb rabbit hole of actually doing AoC as a competition there's a leaderboard for the first people who finish and you have to solve the problem by like 12:00:00.0001 every night if you want to get on it... Humans will always optimize the fun out of anything
|
# ? Dec 3, 2021 18:28 |
|
default behavior of multiple args to "for" in lisp is cartesian product of sequences and you gotta use the loopin dsl to gently caress w things in less circuitous way and its pissing me off
|
# ? Dec 3, 2021 18:30 |
Carthag Tuek posted:i dont like that you have to use is for null, its annoying imo it’s one of the redeeming features of the null, as my personal preference would be to have N/A (“undefined”) and NULL (“void”) separately. but i also had to supervise a dozen non-technical sql query authors for a few years, so ive seen bad and then some
|
|
# ? Dec 3, 2021 18:44 |
|
cinci zoo sniper posted:imo it’s one of the redeeming features of the null, as my personal preference would be to have N/A (“undefined”) and NULL (“void”) separately. but i also had to supervise a dozen non-technical sql query authors for a few years, so ive seen bad and then some tbh its a small annoyance, i think i just dont get why it has to be that way null is a keyword right? is there a case where "select * from table where x = null" shouldn't produce the same results as "select * from table where x is null"?
|
# ? Dec 3, 2021 19:18 |
|
Carthag Tuek posted:tbh its a small annoyance, i think i just dont get why it has to be that way null has no value, therefore null == null is false.
|
# ? Dec 3, 2021 19:20 |
|
MrQueasy posted:null has no value, therefore null == null is false. i know it makes logical sense, but is there a scenario where it makes practical sense is what im asking like i always have to work around it, i cant think of a db structure where null did not equal null (for the same column type)
|
# ? Dec 3, 2021 19:30 |
|
Carthag Tuek posted:tbh its a small annoyance, i think i just dont get why it has to be that way i think the issues arise when you're doing the opposite of your examples; i.e. just going about your day joining existing (hopefully but not necessarily) non-null things to other things. If null=null then if you have nulls in the join columns on both sides they explode in a really gross and bad way
|
# ? Dec 3, 2021 19:31 |
|
i tried to think up an example but the timer went off for my quiches so that exercise is left to the reader
|
# ? Dec 3, 2021 19:33 |
|
I love SQL but for the sake of new people I'd love it if the order of statements was changed and select went at the end
|
# ? Dec 3, 2021 21:00 |
|
the good thing about sql is that it uasnt fundamentally changed in 50 years cuz problem didnt change the bad thing is not composable. fix w query gen imo
|
# ? Dec 3, 2021 21:06 |
|
pointsofdata posted:I love SQL but for the sake of new people I'd love it if the order of statements was changed and select went at the end yeah tbh im not a fan of delete & update without a clause just going ahead. id like if you had to go "yes everything" in the statement at least as a config option
|
# ? Dec 3, 2021 21:09 |
|
Carthag Tuek posted:i know it makes logical sense, but is there a scenario where it makes practical sense is what im asking NihilCredo posted:ansi nulls are correct for the former, because you don't want 'a.foo = b.foo' to return true when there is no match in either a or b
|
# ? Dec 3, 2021 21:14 |
|
maybe if your data isnt poo poo, that wont be a prroblem? im too buzzed to think of an example where i dont want null = null can you do
|
# ? Dec 3, 2021 21:19 |
|
Carthag Tuek posted:maybe if your data isnt poo poo, that wont be a prroblem? it's less null == null and more null <> null and null < null.
|
# ? Dec 3, 2021 21:24 |
|
MrQueasy posted:it's less null == null and more null <> null and null < null. yeah those make sense i guess i just want an option where = includes null y know for my bespoke queries that i spend a lot of time on making sure wont gently caress me, but i also dont want to spend the time typing is instead of = give me a footgun pls
|
# ? Dec 3, 2021 21:27 |
Carthag Tuek posted:maybe if your data isnt poo poo, that wont be a prroblem? yeah if you’re unemployed and have no real world data at all it’s not a problem
|
|
# ? Dec 3, 2021 21:28 |
|
cinci zoo sniper posted:yeah if you’re unemployed and have no real world data at all it’s not a problem just gimme write on prod, whats the worst that could happen?
|
# ? Dec 3, 2021 21:44 |
|
Carthag Tuek posted:just gimme write on prod, whats the worst that could happen? I had write access on prod when I first joined years and years ago and was young and dumb. all the passwords were the same for everything (username backwards). I remember sitting with a Dev and saying "there's a missing criteria on this join that's why this stuff fails" and him going "this here? *F5* fixed" and thinking "oh cool we can do that?" I only hosed up once and deleted some minor historic export table which I then manually restored before anyone noticed (nobody cared anyway), that was a scary enough experience I was glad it happened with something minor. pretty sure I was still cowboying in updates/inserts in like 2011 before anyone started actually looking at what we were doing. Now it's all controlled but the people applying the controls know gently caress all about what were doing so really have no idea if we're just writing a messed up delete statement and wrecking poo poo.
|
# ? Dec 3, 2021 22:19 |
|
NihilCredo posted:no null vs undefined distinction only a javascript programmer could possibly think this is something you'd WANT to have
|
# ? Dec 3, 2021 22:34 |
|
Carthag Tuek posted:yeah those make sense instead of "=" you can write "is not distinct from" and get the behavior you want
|
# ? Dec 3, 2021 22:35 |
|
this post was in the chamber so im posting it: (i actually do have write in prod now, but also its a 20 year old db that moves fairly slow so even a weeks loss is less than 1k insert/updates) one time 15 years ago, i got full access immediately, cause this guy i never met vouched for me*. i hosed up a lot with that db... but the good thing is that the system itself was brittle as gently caress and there was no logging and nobody got hurt except multinationals * i was doing 1st line support & he had my email from talking about the file format from a 90s game online. idk** why he just quit, but he reached out and i apprenticed for a bit and took over his job. then he went to work for spies or something ** the bosses were garbage. very cargo cult startup energy
|
# ? Dec 3, 2021 22:38 |
|
|
# ? May 9, 2024 03:06 |
DELETE CASCADE posted:only a javascript programmer could possibly think this is something you'd WANT to have when i was working on data quality management topics a few jobs ago, i ended up standardising everything around 3 types of nulls - “this is not possible”, “explicitly void”, and “other”
|
|
# ? Dec 3, 2021 23:13 |