DB THREAD
Images are sometimes not shown due to bandwidth/network limitations. Refreshing the page usually helps.
You are currently reading a thread in /g/ - Technology
You are currently reading a thread in /g/ - Technology
what's /g/ preferred db/keystore?
My list
>MySQL
>MongoDB
>Redis
I use cockmonglerdb as well
>>52085983
Oracle SQL is probably the best
What would I need a keystore for? Genuinely want to know as I'm writing a niche social network app and it might come in handy but I don't know the context
>>52086896
caching
>>52086896
if you don't already know what it is, don't bother using it. chances are you just need good ol' MySQL or PG.
>>52086914
But what sort of data would be useful to cache? I guess it's a chicken-egg thing, I haven't completed my API yet so it's probably something to research after that's done
Microsoft SQL Server :^)
Tried mongodb for handling logging requests and output from a web service at work, due to lack of output schema other than "it's gonna be json ;)"
Sadly realised that the we should've gone with the usual suspect instead (ms sql/oracle/mysql) etc.
Oh well, life lessons.
>>52086952
if you fetch data from a service, cache it locally in a key store to avoid calling the service again... Also it's faster than a db query cause it's loaded in memory.
>>52085983
>db/keystore
what the fuck
those are not one and the same
>>52087353
>I can't read between the lines
>I probably own a mac
If you're looking for a good oldschool solution then >>52086884 is definitely your best bet
You need to be at least semi-competent in relational database design though
But MongoDB and ORMs are what the world is moving towards, with good reason
>>52087391
fuck you
PostgreSQL > MySQL for RDBMS.
Redis cool. Haven't touched mongoDB.
>>52087492
Anything is better than MySQL desu
only phpniggers praise it
>>52087492
>>52087535
pls explain why
>>52087449
ORMs are pure shit and slow down any application.
>B-but muh object methods
God I hate ORMs fags
For 99.999% of the population, an RDBMS is more than adequate; they've been around for decades and are battle tested. MariaDB is good for the majority, PostgreSQL for more advanced setups if the requirements dictate the need for such capabilities, and Oracle if you work for a cucked company that doesn't care about spending $17,500 for a database with a feature set that you may or may not (probably the latter) need.
I'm willing to bet that every application that uses MongoDB could've easily gotten away with using MariaDB without even getting close to taxing its capabilities, and would've been better off as a result.
Add Redis later only if your database is being hit heavy enough to warrant the extra complexity of caching in order to take some burden off of the database.
>>52087449
Worst post in this thread.
>>52087596
They're not perfect, but they migrate DB management entirely to the application layer, which is definitely a big point in their favour
always relevant
https://www.youtube.com/watch?v=b2F-DItXtZs
(Stupid) question! So far I only used MySQL. How is Redis different from a MySQL MEMORY table?
>>52087642
>what is object-relational impedance mismatch
>what is SQL being pure shit and not being close to anything described in the manifestos
if SQL was like D, there'd be some weight to this post
>>52087722
>babby's first web app
>>52087659
ORMs remodel data into objects, so say, every time you query, instead of just giving you, say, a json or xml output, a truckload of shit is recreated in memory so some stupid fag can access some method at object level in the app. it's heavy on memory and CPU. Every app i've worked on that used ORMs where slow and buggy as hell and changing one table would throw you into a rebuild nightmare.
>>52087754
>babby's first shitpost
>>52087763
Don't use a shit ORM and optimize your queries
>>52087708
MySQL memory tables: performance is constrained by contention resulting from single-thread execution and table lock overhead when processing updates. This limits scalability when load increases, particularly for statement mixes that include writes.
>>52087708
Redis is usually used as a stand alone solution for caching purposes. If that's all you need, then using MySQL for only that purpose would be overkill. Redis is a simpler solution.
>>52087827
>>52087848
Thanks senpais. Sounds good, I think I'll look into Redis.
>>52085983
postgre or mongodb for javascript projects
>>52087903
Not to mention that Redis was specifically designed for this purpose and will work better than MySQL as >>52087827 notes.
https://www.youtube.com/watch?v=b2F-DItXtZs
>>52085983
>mongoDB BTFO
http://arstechnica.com/security/2015/12/13-million-mackeeper-users-exposed-after-mongodb-door-was-left-open/
Friendly reminder to RTFM when you use a database. Check yo self before you wreck yo self faggots
>>52088690
this is the plague of HipJSter
>muh npm installs
>muh github
>>52088690
tbqh senpai, mongo doesn't come with an admin user defined by default, which is pretty shitty if you ask me
>>52089998
>defined by default
raw text files
>>52085983
Riak.
Sqlite3 only cuckolds and enterprises use anything else.
>mongo
>the udp of databases
>good
kek
postgres or bust
>>52085983
I don't understand how db like mongocb work. Someone has some literature that explain it ?
>>52091554
Look around, it has tons of problems. It is error prone and by default was set to silence all errors and to not confirm if data was successfully written, issues with random data loss, it's slower than advertised, and forces certain schema on the user.
>>52087661
kek
>>52091644
just set the write concern option on
>>52091522
>>52091644
sounds like a bunch of lies made up by sql programmers who want to keep their jobs
>>52091554
it's a document based database. Instead of having multiple tables, you have a collection of documents. They can have different or common properties, such as
>john.age = 20
>john.sex = m
>jack.age = 10
>jack.sucks = dicks.
You can query the db to return only documents who have the age property. For more complex operations, you do a "map reduce", which is basically going through all the documents your query returns and reduce the number of documents by cycling through them with a javascript function while passing along some data passed from one to another.
MySQL master race.
>>52093168
MySQL is pleb race. Oracle or Postgres, or fuck off.
Mysql mostly.
Redis for my caching
Mongodb sometimes depends on the data.
>>52093168
>>52093277
Even if it's not oracle/postgres, at least use MariaDB instead of MySQL.
>>52093672
please explain why
mongoloidDB
PostGres and Redis
I haven't worked on anything where mongoloidDB made sense, but it looks pretty cool I guess.
>>52086896
Redis specifically has comparable (slightly slower, ignoring your language bindings) performance to memcached but efficient support for a number of data structures that often enable a better caching strategy. I wrote a correct expiring ring cache with redis in like 30 lines of python that outperforms simple key-expiration caching really well for the workload I implemented it for. Also it works quite well as a language/OS-independent IPC mechanism. Throwing around byte arrays and trusting the other side to decode is a poor idea on the whole, both because it's not very expressive and because it creates an unequal relationship between platforms (the decoder is almost always the encoder to the target's data representation schema as well because it's the only way to get one way coupling vs. two way).
