|
DONT THREAD ON ME posted:i get it but you can build a thing in postgres with a jsonb column (or with a schema because it's really not that hard) without caring about it and it will take the same amount of effort, work better, and you wont be hosed when your bullshit thing suddenly becomes important. oh yeah i'm talking about schemaless in general, not mongo i don't think i've ever run into someone who's run a real project on mongo who didn't hate it
|
# ? Jan 15, 2019 02:52 |
|
|
# ? May 3, 2024 11:03 |
|
Rex-Goliath posted:oh yeah i'm talking about schemaless in general, not mongo oh yeah that's reasonable. be pragmatic. you're definitely better off going with post gres though, then you only need to do a migration if you ever need to switch to a real schema.
|
# ? Jan 15, 2019 02:55 |
|
Rex-Goliath posted:lol at some of you being beaten into submission and feeling like you just gotta have a schema because what’s the big deal?? once you’ve built something without having to care about it you’ll understand the appeal yeah, you don't build it "properly" because you're not a subject matter expert on the underlying database engine, so your site performs terribly and then obama hires a bunch of google employees to fix it.
|
# ? Jan 15, 2019 03:00 |
|
lol
|
# ? Jan 15, 2019 03:10 |
|
Fiedler posted:yeah, you don't build it "properly" because you're not a subject matter expert on the underlying database engine, so your site performs terribly and then obama hires a bunch of google employees to fix it. tell me how they hurt you
|
# ? Jan 15, 2019 03:51 |
|
i love to have my schemas defined implicitly by the code using the data so that i can never be quite sure what fields should be present
|
# ? Jan 15, 2019 05:43 |
|
Skim Milk posted:necc0 is now bad database shaggar
|
# ? Jan 15, 2019 05:52 |
|
Plorkyeran posted:i love to have my schemas defined implicitly by the code using the data so that i can never be quite sure what fields should be present this but unironically
|
# ? Jan 15, 2019 06:22 |
|
best thing about having no schema is the abundance of null-checks that happen in code
|
# ? Jan 15, 2019 07:18 |
KOTEX GOD OF BLOOD posted:liking things is illegal in yospos if liking you are posting is wrong, i dont want to be right
|
|
# ? Jan 15, 2019 07:20 |
giant cock man, hows the new crib so far?
|
|
# ? Jan 15, 2019 07:21 |
|
i liked couchdb
|
# ? Jan 15, 2019 07:29 |
|
Who loving cares what database your dumb poo poo is using just trust the ORM and ship it
|
# ? Jan 15, 2019 08:17 |
|
in my brief attempt to understand databases I *think* I understand a relational database but I can't for the life of my conceptualise an entirely different concept, unless that concept is basically just a filesystem without a filesystem, ie, data/documents with names that you can access building a dynamic web page when I only put in an hour a week is hard, and have to learn python, postgres, django, javascript and html. I think. installing enough poo poo on my macpro to get it all running by installing pip and poo poo takes long enough as it is. I have a webpage that currently goes to one of two pages depending on if you type in one of two urls... and I think learning how to database is going to be key to making my website worthwhile. but also is the javascript and html. anyway lol its a lot to learn when I can never be hosed doing anything on it
|
# ? Jan 15, 2019 09:02 |
|
its gonna be good tho belieb u me
|
# ? Jan 15, 2019 09:03 |
|
echinopsis posted:in my brief attempt to understand databases I *think* I understand a relational database but I can't for the life of my conceptualise an entirely different concept, unless that concept is basically just a filesystem without a filesystem, ie, data/documents with names that you can access that's what it is. i mean it's more than that, better performance and querying features, the documents arent usually files but usually some structured data, it's usually stored partially in memory, etc. the same thing confused me at first too.
|
# ? Jan 15, 2019 09:07 |
|
Didn't Postgres with a PK, jsonb table outperform mongo when you turned off all the synchronous commit and fsync safety and when you enabled all that pesky "database" stuff in mongo it was slower than Postgres's defaults?
|
# ? Jan 15, 2019 12:29 |
|
SpaceAceJase posted:Who loving cares what database your dumb poo poo is using just trust the ORM and ship it yeah, but my impotence mismatch.
|
# ? Jan 15, 2019 13:05 |
|
fart simpson posted:this but unironically
|
# ? Jan 15, 2019 14:17 |
|
Skim Milk posted:giant cock man, hows the new crib so far? p good p good. developer is a scatterbrained idiot who i’m gonna have to take to small claims court over some small stuff but overall i’m happy
|
# ? Jan 15, 2019 14:47 |
|
Trabisnikof posted:i liked couchdb its p dece yeah
|
# ? Jan 15, 2019 14:52 |
|
Mongodb should have just sold seats on it's board for $30 or $50k.
|
# ? Jan 15, 2019 17:04 |
|
Rex-Goliath posted:lol at some of you being beaten into submission and feeling like you just gotta have a schema because what’s the big deal?? once you’ve built something without having to care about it you’ll understand the appeal i have done what you say before and i did not understand the appeal. if you want to prototype something schemaless as a proof of concept, where you only care about the happy path through your code, then sure i guess that's ok. but as soon as you have a persistent datastore, why would you NOT want a schema? unless you're doing free text search, your program is gonna care about the structure of your data, and the data will probably outlive your code, so why wouldn't you want your database to do all it can to help? you're implicitly defining a schema regardless... sql ddl is not hard
|
# ? Jan 15, 2019 18:08 |
|
DELETE CASCADE posted:i don't understand why anyone would use mongodb or even want a document store. your poo poo's relational, yes it is, i promise, look harder at it, or have someone else design the schema if your brain is too broken from staring at json all day. then you get to use sql and a stable rdbms and live happily ever after I’ve worked a year and a half on a project with a in house json document store and all I can say is blech bring back relational databases
|
# ? Jan 15, 2019 18:14 |
|
There is always a schema. Whether this schema explicitly exists in a file, or is implicitly constructed by the fields that the code looks for is up to you.
|
# ? Jan 15, 2019 18:36 |
|
people who use document stores for things that aren't documents are lovely people and give software development a worse name that it already has. get out of my field, dummies.
|
# ? Jan 15, 2019 18:56 |
|
can't we just blame mongos period of popularity on front-enders going "full stack" aka we need jabascript end-to-end
|
# ? Jan 15, 2019 19:33 |
|
DELETE CASCADE posted:i have done what you say before and i did not understand the appeal. if you want to prototype something schemaless as a proof of concept, where you only care about the happy path through your code, then sure i guess that's ok. but as soon as you have a persistent datastore, why would you NOT want a schema? unless you're doing free text search, your program is gonna care about the structure of your data, and the data will probably outlive your code, so why wouldn't you want your database to do all it can to help? you're implicitly defining a schema regardless... sql ddl is not hard if your schema is defined in code why are you redundantly defining it again? isn’t that what unit/integration tests are for?
|
# ? Jan 15, 2019 19:54 |
|
Rex-Goliath posted:if your schema is defined in code why are you redundantly defining it again? isn’t that what unit/integration tests are for? serious answer: because usually it's unclear and fragile in code, split across lots of different places. explicit schemas define it in one unarguable place and unambiguously non-serious answer: unit tests are redundant with the code, why bother with them. if there's an issue your users will tell you
|
# ? Jan 15, 2019 21:19 |
|
if it’s something critical like a primary key or whatever you can enable schema validation on any elements which shouldn’t change or have to follow certain guidelines. with that other systems are still free to write additional information should it become necessary without all the front work being necessary
|
# ? Jan 15, 2019 21:22 |
|
Rex-Goliath posted:if your schema is defined in code why are you redundantly defining it again? isn’t that what unit/integration tests are for? I feel like this is the same argument I have with dynamic typing advocates. Sure, I can write unit tests to verify at runtime that every value has the type I think it does. But why not let the type system do that for me? It’s less work, less error prone, easier to maintain, easier to understand and communicate, and allows for optimizations that penetrate even through layers of code I don’t control. It’s not redundant because I don’t write unit tests to verify that an int is an int, or in this case that a column has the proper type or foreign key relationship. I’d also reiterate what I said in the post you quoted: the data Is not owned by my code, simply processed by it. Other codebases belonging to different teams also manipulate it, and I’d like to enforce its integrity in the same central location where it is stored. That way I don’t have to run my team’s test suite only to find that the data has been corrupted by someone else’s code. In fact, you could throw my code away, and the data will remain just fine for the next project that comes up, documented and enforced by the schema.
|
# ? Jan 15, 2019 21:28 |
|
"that just sounds like a database with extra steps"
|
# ? Jan 15, 2019 21:30 |
|
if you’re concerned about that you can still enable schema validations in document stores. hard schemas don’t only exist in relational dbs
|
# ? Jan 15, 2019 21:31 |
|
Rex-Goliath posted:if you’re concerned about that you can still enable schema validations in document stores. hard schemas don’t only exist in relational dbs great now i'm using a document db AND i have to define a schema? why not just expend the extra 30 minutes of effort and define a schema that will make my life and everyone elses life easier? DONT THREAD ON ME fucked around with this message at 21:48 on Jan 15, 2019 |
# ? Jan 15, 2019 21:44 |
|
Isn't storing json far less storage efficient because it stores the fields with every document
|
# ? Jan 15, 2019 21:46 |
|
because you can store full entities which can be as complex as you want without having to deal with setting up all the tables and relationships first. if you’re using xml you can also use namespaces to version your schema which allows you to refractor data on a production system with zero downtime or gymnastics which is nice
|
# ? Jan 15, 2019 21:48 |
|
abigserve posted:Isn't storing json far less storage efficient because it stores the fields with every document depends on the db. the smart ones split the docs apart into a series of trees and pointers which allows for heavy overlapping. there’s probably a lot of real dumb implementations out there too
|
# ? Jan 15, 2019 21:50 |
|
Rex-Goliath posted:because you can store full entities which can be as complex as you want without having to deal with setting up all the tables and relationships first. if your data is relational this ends up turning into a pile of spaghetti bullshit extremely fast. i just went through 2 years of dealing with this. never again. if your data isn't relational then fine, i guess. e: sorry for being a dick schemaless ruined my life for the last 2 years. DONT THREAD ON ME fucked around with this message at 21:56 on Jan 15, 2019 |
# ? Jan 15, 2019 21:51 |
|
yeah i’m literally a nosql consultant i’ve seen my fair share of horror stories. it’s absolutely not a replacement for considerate design and general best practices and tons of shops treat it that way. just because you don’t have to have a schema doesn’t mean it’s not important for certain entities if you know what you’re doing though you can build some pretty crazy poo poo in months that would be effectively impossible for a relational system
|
# ? Jan 15, 2019 22:03 |
|
|
# ? May 3, 2024 11:03 |
|
Rex-Goliath posted:yeah i’m literally a nosql consultant sorry about your loss
|
# ? Jan 15, 2019 22:07 |