HOW do we KILL this slow piece of shit garbage from the 70s

Frick this mickey mouse ass language you can't even do anything in it

Shopping Cart Returner Shirt $21.68

Black Rifle Cuck Company, Conservative Humor Shirt $21.68

Shopping Cart Returner Shirt $21.68

  1. 1 month ago
    Anonymous

    SQL isn't slow moron. Just write better queries.

    • 1 month ago
      Anonymous

      Because there's nothing else to compare against! Why can't I just call the C modules directly and get my data from there?

      • 1 month ago
        Anonymous

        Dumbass.

        • 1 month ago
          Anonymous

          SQL is inefficient garbage then, thanks for confirming.

      • 1 month ago
        Anonymous

        Did you look up why you can't before you made this thread? Because you would know why

        Dumbass.

        He might be mentally challenged, ease up. (morons are so dumb they like being called dumb these days)

        • 1 month ago
          Anonymous

          Another idiot who never questions anything in life and just accepts whatever his masters give him.

      • 1 month ago
        Anonymous

        I'm amazed that you think interpreting a SQL query is what's taking up the majority of a queries run time.

        • 1 month ago
          Anonymous

          And how much faster is reading the same file directly from disk? Why do SQLoids never ask this question?

          • 1 month ago
            Anonymous

            SQL doesn't require reading a file from disk. What the frick are you actually trying to say?

          • 1 month ago
            Anonymous

            Ok, you're moronic and obtuse, congratulations. You can't even entertain a simple benchmark scenario.

          • 1 month ago
            Anonymous

            What?

          • 1 month ago
            Anonymous

            In 99% of real life scenarios, when a SQL server executes a query, something, somewhere is getting read from a "file" on something represented as a block device, whether it's contiguous RAM, a LUN, a local disk, etc.

          • 1 month ago
            Anonymous

            Make a 100 GB file and see how long it takes to find something in it.

          • 1 month ago
            Anonymous

            see this is why i dont use big hard drives. fricking morons bragging about their 24TB drives. I have things to do, i'm not going to wait for my pc to search through LITERAL TERABYTES everytime i need something from my hard drive

          • 1 month ago
            Anonymous

            my raidz2 array of 20*12TB hard drives is faster than your sata SSD so um skill issue?

          • 1 month ago
            Anonymous

            It's not. Even a single SATA SSD destroys it in IOPS. A HDD is at most, and I'm being generous here, 500 4K QD1 IOPS. Increasing QD decreases IOPS for HDDs.
            Shitsung 870 EVO SATA does 13000 IOPS QD1, so you'd need 26+ HDDs to even match it in raw performance. For QD32 that single SSD does 98000 IOPS. NVMe goes into millions nowadays.
            Oh and ZFS RAID-Z decreases IOPS naturally, as soon as you run out of ARC and L2ARC it tanks heavily.

          • 1 month ago
            Anonymous

            you are utterly moronic and have no clue how raid works
            i get sustained reads from my array of 1.2gigaBYTES a second and writes aren't far behind at 1GB/s
            my samsung evo sata SSD caps out at 600MB/s because that's the maximum speed of a sata 3 interface for a single drive. my array is almost twice as fast as a sata 3 ssd and about as fast as a pci gen 2 NVMe ssd. if it was raid 10 it's be about as fast as a gen 3 (3GB/s) and if it was raid 0 it's be faster than gen 4 pci at 6GB/s.

            go look up raidz2 write factor and learn some basic math before you come back

          • 1 month ago
            Anonymous

            >i get sustained reads from my array of 1.2gigaBYTES a second and writes aren't far behind at 1GB/s
            So you're just ignoring the whole IOPS thing, hilarious.
            I have a RAID-Z of 3 Shitsung 15TB U.3 NVMe drives. Do you know what speed I get on them? 32 GB/s write and 40GB/s read, yes just three. Do you know how much IOPS this gives? 55000 at QD2 which is insane for RAID-Z. Oh and it's with ARC at 1GB to measure just the disks.
            >go look up raidz2 write factor and learn some basic math before you come back
            Go look up how RAID-Z scales IOPS.

          • 1 month ago
            Anonymous

            moron. HDDs have terrible seek times, since they have to move the read head and wait until the disk spins to where the data is. SSDs don't have such problems so 1 SSD can destroy your puny raid array in I/O Operations per Second. (IOPS).

          • 1 month ago
            Anonymous

            find a storage room.
            now, pick up a manifest and find a box number, then go grab that box from the indicated shelf. pretty quick to just skim down a list, right?
            try find that box without the manifest. it takes a lot fricking longer now, doesn't it?

          • 1 month ago
            Anonymous

            >And how much faster is reading the same file directly from disk? Why do SQLoids never ask this question?
            It's been answered long ago for many different DB engines. Once you have the indices set up, it takes about the same time to move the data, and less time than a straight filesystem lookup if you've got to do some sort of searching for it.
            Where it really wins is when you'd otherwise have to have hundreds of thousands of little files open (especially for writing reliably). Your OS is not optimized for that (you run out of file handles/descriptors) but having that many rows in your DB is small beans.

        • 1 month ago
          Anonymous

          U mean its not little Swiss gnomes inside of the computer translating SQL to sign language to other gnomes so they can do the required tasks?

      • 1 month ago
        Anonymous

        Filtered

        Filtered

        SQL is inefficient garbage then, thanks for confirming.

        Filtered

        Another idiot who never questions anything in life and just accepts whatever his masters give him.

        Filtered

        And how much faster is reading the same file directly from disk? Why do SQLoids never ask this question?

        Filtered

        Ok, you're moronic and obtuse, congratulations. You can't even entertain a simple benchmark scenario.

        Filtered

        you can learn that shit in an afternoon, do the easy level SQL questions on hackerrank, and use w3schools as a reference, if you need more practice: https://www.sql-practice.com/
        Also here's a good reference for how joins work:
        https://joins.spathon.com/

        You can learn the basics in an afternoon, but writing good sql takes time. But if you're under a few hundred gbs in scale it doesn't fricking matter because sql is that good.

      • 1 month ago
        Anonymous

        >Why can't I just call the C modules directly and get my data from there?

        here OP, I'm actually here to help you:
        https://rocksdb.org/
        or if you're a go homosexual: https://dgraph.io/docs/badger/

        good luck and god speed. don't let them take you down.
        oh, you might also need to understand a few other helpful concepts to accelerate your post-sql escape: https://roaringbitmap.org/

    • 1 month ago
      Anonymous

      Unless you index everything, SQL is O(log(n))
      A .ini file or a hashtable is O(1).

      • 1 month ago
        Anonymous

        >A .ini file
        Not if that .ini file is 2 TB, and you don't know where the info you want is. Unless you have... an... index... to speed it up.
        >or a hashtable is O(1).
        Only if you know the key beforehand, and only if your hashing algorithm works well for your data. Hash tables are routinely used in database engines.

  2. 1 month ago
    Anonymous

    Is there a preferred way to learn SQL?
    I frick around in Excel/Power Query a lot for work, and it occurred to me the other day that- at the point I'm beginning to rely on Table.Join() I might as well skip the middleman.

    • 1 month ago
      Anonymous

      you can learn that shit in an afternoon, do the easy level SQL questions on hackerrank, and use w3schools as a reference, if you need more practice: https://www.sql-practice.com/
      Also here's a good reference for how joins work:
      https://joins.spathon.com/

      • 1 month ago
        Anonymous

        Sorry for the late reply, but thank you; I mainly learn through repetition, and don't have access to the database at work, so the links are really appreciated.

      • 1 month ago
        Anonymous

        >https://joins.spathon.com/
        i wish i knew this site before...

  3. 1 month ago
    Anonymous

    SQL is kind of an ass language, but crucially it's not quite bad enough for anyone to feel any urgency to move away from it. Like it only really starts sucking when you have a select inside a select inside a select inside a select inside a select, but those queries aren't too common.

    • 1 month ago
      Anonymous

      Those queries are written by poos or drag and drop GUI type tools. Use CTEs or temp tables. beware of CTEs in certain languages because if you reference them multiple times it actually executes them multiple times instead of caching the result.

    • 1 month ago
      Anonymous

      Do you not realize they made name-value pair 'databases'? I was of the very first if not the very first to suggest that as an option for circumstances that did not require complex queries and people said "solution for problem that does not exist". Well NoSQL databases exist because not all situations require a SQL interpretor and overlay to get at data.

      So yes 'geniuses' have realized moving away from it when necessary can boost productivity. But the only people who would think of abandoning SQL altogether have never done anything beyond basic bullsht.

  4. 1 month ago
    Anonymous

    your right we should just be using javascript and json. sql is boomer garbage created to keep us poor

    • 1 month ago
      Anonymous

      If you ever worked with anything beyond the most basic kiddie project you would know why SQL is a scripting language to access and managed structured data instead of *writing an entire program from scratch everytime you want to access or manage that data. SQL exists for the same reason people drop lua or javascript or some other scripting into a specific doman (such as lua + nginx). SQL is a domain-specific scripting language.

      Whining about SQL is like whining about using HTML and Javascript and saying "Waaaaaah, why dont we just write everything in C or Assembly?"

  5. 1 month ago
    Anonymous

    >can't even do anything in it
    https://demozoo.org/productions/268459/

    • 1 month ago
      Anonymous
  6. 1 month ago
    Anonymous

    I'd not kill myself if I was as smart as OP.

  7. 1 month ago
    Anonymous

    zoomies do be moronic

  8. 1 month ago
    Anonymous

    >from the 70s
    it must be pretty fast if it could run on 70s computers

  9. 1 month ago
    Anonymous
  10. 1 month ago
    Anonymous

    op is an actual moron, sql is fine oracle database is shit it still used sql but nothing ever works right

  11. 1 month ago
    Anonymous

    I love how SQL filters dumb jeet morons like you. Enjoy unemployment.

  12. 1 month ago
    Anonymous

    lol zoomers can’t into SQL. You think that’s slow? Try poorly designing a mongodb app and tell me about slow.

  13. 1 month ago
    Anonymous

    Postgres is faster than any other DB

    • 1 month ago
      Anonymous

      No it isnt, lmao

      • 1 month ago
        Anonymous

        https://www.techempower.com/benchmarks/#hw=ph&test=query&section=data-r22

        • 1 month ago
          Anonymous

          >axum [mongodb raw] 6,727
          >axum [postgresql] 32,591
          damb. but this doesn't really say much about other DBs. only flavors they have are: postgres, mysql or mongoloid.

    • 1 month ago
      Anonymous

      MongoDB is Web scale, therefore faster

  14. 1 month ago
    Anonymous

    >slow
    >can't even do anything
    obvious skill issue

  15. 1 month ago
    Anonymous

    Too bad object databases never gained widespread acceptance.

  16. 1 month ago
    Anonymous

    I love how SQL filters proceduralbrained cavemen.
    Go back to your C fizzbuzz, morons.

  17. 1 month ago
    Anonymous

    I LOVE how everybody in this thread completely IGNORES the fact that they live in a walled garden full of GARBAGE.
    SQLoids, LISTEN UP!!!! YOU ARE IN A BUBBLE, YOU ARE IN A MATRIX. WAKE UP.
    SQL is a bubble the same way Javascript is. You are under the influence of Stockholm syndrome and you can't see how dire your situation is. You DON'T need a JIT parser to read bytes from a file, I KNOW, I KNOW, it's a crazy thought isn't it? But it's true, there is an entire world outside the SQL bubble, you just can't see it.

    • 1 month ago
      Anonymous

      SQL is ok. I'm pretty comfortable with storage systems in general so I usually shirk sql, but most modern sql driven engines are feature rich and it's easier to just conform to their model of data storage and query to get these features. You'll end up reinventing a lot of shit if you go full KV only chad.

  18. 1 month ago
    Anonymous

    just use T-SQL then.

  19. 1 month ago
    Anonymous

    moron OP, moron opinion, moron thread.

  20. 1 month ago
    Anonymous

    I work for a company that makes medical devices (digital pathology). we work with petabytes of data on the daily and PostgreSQL suits us just fine. if you need more scale than that you can use whatever gay ass zoomer mongodb you want

    • 1 month ago
      Anonymous

      >works with petabytes of data
      >zero reading comprehension
      The SQL JIT walled garden has really rotted your brain. Please don't kill any patients.

      • 1 month ago
        Anonymous

        so how does your mongodb shit not being A.C.I.D a good thing mister junior dev failing at his company to solve the sql task his senior gave him?

        • 1 month ago
          Anonymous

          I haven't mentioned NoSQL or Mongo anywhere in this thread. The third post in this thread says exactly what I want to achieve.

      • 1 month ago
        Anonymous

        have you ever heard of an ORM moron kun? this is how i query from a database:
        [Code]

        var result = ctx.Users
        .Where(user => user.age >= 18)
        .Select(user => user.name)
        .Order();
        [/Code]

        • 1 month ago
          Anonymous

          I DON'T WANNA GO THROUGH THE DAMN JIT. I KNOW EXACTLY WHAT DATA I HAVE AND HOW I AM GOING TO READ IT.

          • 1 month ago
            Anonymous

            then use csv files Black person they are basically your tables. though i guarantee you it will be much slower than "le heckin evilerino jitted" db

          • 1 month ago
            Anonymous

            As soon as his data differs by one byte per "column" it's going to fall apart. The more differences between rows the more reinventing of the RDBMS he's going to do, badly.

          • 1 month ago
            Anonymous

            As soon as his data differs by one byte per "column" it's going to fall apart. The more differences between rows the more reinventing of the RDBMS he's going to do, badly.

            okay, so where do you need SQL in that exactly?

            I want all the guarantees of a database without all the garbage legacy moronation. Do I have to spell it out for you?

          • 1 month ago
            Anonymous

            >garbage legacy moronation
            such as? yes, you do need to spell it out.

          • 1 month ago
            Anonymous

            Having a JIT and not being able to write to the tables and indices using plain C.

          • 1 month ago
            Anonymous

            >what is libpq
            >what is libmysqlclient
            >what is PostgreSQL LLVM-based JIT

          • 1 month ago
            Anonymous
          • 1 month ago
            Anonymous

            Very nice self-portrait. Now explain what you mean exactly.
            You can use the JIT or you can not use it at all.
            Every DBMS has a C library.

          • 1 month ago
            Anonymous

            You fat fricking moron. What do you think you're sending through the socket and what do you think the database does with it?

          • 1 month ago
            Anonymous

            C libraries also make you write sql. you could either write a 1000 functions called GetTable(), GetColumn() and so on and so on and you would have to manually then stitch together this fricking mess into something usable OR you could just script the logic(sql) and get the data in the format you want in its final form from the beginning.

          • 1 month ago
            Anonymous

            You fat fricking moron. What do you think you're sending through the socket and what do you think the database does with it?

            Of course they do use SQL. How else do you imagine you would communicate with it?
            If you want to read the raw data files then just look what the engines do, they are after all written in C and open source. However there is no reason to do this because in order to use that data you'd have to basically reimplement the SQL parsing engine.
            Oh and RDBMS' using JIT are a relatively new thing. The traditional way is to interpret and cache, so calling JIT "legacy" is daft.

          • 1 month ago
            Anonymous

            SQL has nothing to do with this though. It's literally a declarative language, so JIT can only be beneficial.
            You're barking at the wrong tree.

          • 1 month ago
            Anonymous

            1, use an ORM. this way you dont have to write sql and you get all the benefits of a db with the only downside being its slower than handwriting the sql

            2, use a nosql database (they are all garbage in my humble opinion)

            3, suck it up and write sql

            now for the reason why you are not allowed to just start reading/writing bytes randomly: databases HAVE TO manage who and how can modify data to keep data integrity if db's just gave you a pointer and allowed you to start writing whatever you want you could easily overwrite other memory. not to mention no db's would be able to be ACID compliant anymore since how are they supposed to do transactions while you have a pointer randomly writing shit all over the place while the db is trying to do its stuff.

          • 1 month ago
            Anonymous

            They could prevent bad behavior by requesting a proof that the C program will conform to whatever guidelines are required by the db along with the program.

    • 1 month ago
      Anonymous

      >if you need more scale than that you can use whatever gay ass zoomer mongodb you want
      mongo can't handle this scale
      facebook skipped mongo and went to HBase instead, then to MyRocks, which is a RocksDB implementation on MySQL 5.x engine, and finally moved to MySQL 8.x
      and they handle more than peta- scales

  21. 1 month ago
    Anonymous

    I'm using Json for a simple database and its much less of a pain to use than sql. Still haven't run into the limits.

    • 1 month ago
      Anonymous

      Fortunately when you do hit a limit you can easily move your data to modern databases, since they handle JSON natively.

  22. 1 month ago
    Anonymous

    >you can't even do anything in it
    (You) can't. I can. Skill issue.

  23. 1 month ago
    Anonymous

    we won't
    no one in their right mind should try to take on the work people have put into SQL engines, literally 50+ years of concurrency and ACID transactions you do not have to waste time on, but hey, maybe you should take a look at mongo if you want not to use it

    • 1 month ago
      Anonymous

      I don't want to use Mongo. I don't give a shit about DB wars. I only want to read bytes from a file.

      • 1 month ago
        Anonymous

        fopen("./gay_op.txt", "rb");

      • 1 month ago
        Anonymous

        okay, so where do you need SQL in that exactly?

  24. 1 month ago
    Anonymous

    t. literal moron

  25. 1 month ago
    Anonymous

    this is what being a junior dev/university student watching jon blow/casey videos does to your brain

    • 1 month ago
      Anonymous

      Every single site they use is running on SQL, but their projects are too good for it, apparently.

  26. 1 month ago
    Anonymous

    IQfy - People who never question anything in life and only accept what they've been given and never try to improve on it, the board.

    • 1 month ago
      Anonymous

      homie we told you how it literally cannot be done in any other way without reinventing the whole thing or making up something shittier. AS IN THE FRICKING LAWS OF THE UNIVERSE IMPOSSIBLE. if you really just want a pointer to data on a disk use a BINARY FILE. just make a ./data.dat file and open it with fopen() and read/write whatever the goddamn frick you want "fast" (fun fact sql db's will be much faster). or use a onions nosql db like mongodb

      • 1 month ago
        Anonymous

        you could mmap it instead, but it would only make a little difference

  27. 1 month ago
    Anonymous

    The point of SQL is reducing the amount of data that you need to transfer between the DB server and the machine running the application when you have the DB and the application running on different machines.
    If you have to retrieve every value before requesting the key from the other table it's going to be slower than sending the relations and letting the DB figure out the final value by itself.
    Another advantage is letting the DB figure out how much of the data to keep in memory vs retrieving from disk.

Your email address will not be published. Required fields are marked *