❓
물음표살인마 블로그
  • README
  • ALGORITHM
    • Sieve of Eratosthenes
    • Round Up
    • Binary Search
    • Union Find
    • Sorting Array
    • Lcm, Gcd
  • TechTalk Review
    • Template
  • Books
    • CS Note for Interview
      • Ch1. Design Pattern & Programming paradigm
        • 1.1.1 Singleton Pattern
        • 1.1.2 Factory Pattern
        • 1.1.4 Observer Pattern
        • 1.1.5 Proxty Pattern & Proxy Server
        • 1.1.8 Model-View-Controller Pattern
        • 1.2.1 Declarative and Functional Programming
        • 1.2.2 Object Oriented Programming
      • Ch2. Network
        • 2.2.1 TCP/IP Four-Layer Model
        • 2.2.1-1 TCP 3, 4 way handshake
        • 2.3 Network Devices L4, L7
        • 2.4.1 ARP, RARP
        • 2.4.2 Hop By Hop Communication
        • 2.4.3 IP Addressing Scheme
      • Ch3. Operating System
        • 3.1.1 Roles and Structure of Operating Systems
        • 3.2.1 Memory Hierarchy
        • 3.2.2 Memory Management
        • 3.3.1 Processes and Compilation Process
        • 3.3.3 Memory Structure of a Process
        • 3.3.4 Process Control Block (PCB)
        • 3.3.5 Multiprocessing
        • 3.3.6 Threads and Multithreading
        • 3.3.7 Shared Resources and Critical Sections
        • 3.3.8 Deadlock
        • 3.4 CPU Scheduling Algorithm
      • Ch4. Database
        • 4.1 Database Basic
        • 4.2 Normalization
        • 4.3 Transaction and Integrity
        • 4.4 Types of Databases
        • 4.5 Indexes
        • 4.6 Types of Joins
        • 4.7 Principles of Joins
      • Ch5. Data Structure
    • Learning the Basics of Large-Scale System Design through Virtual Interview Cases
      • 1. Scalability based on user counts(1/2)
      • 1. Scalability based on user counts(2/2)
      • 2.Back-of-the-envelope estimation
      • 3. Strategies for System Design Interviews
      • 4. Rate Limiter
      • 5. Consistent Hashing
      • 6. Key-Value System Design
      • 7. Designing a Unique ID Generator for Distributed Systems
      • 8. Designing a URL Shortener
      • 9. Designing a Web Crawler
      • 10. Notification System Design
      • 11. Designing a News Feed System
      • 12. Chat System Design
      • 13. AutoComplete
      • 14. Design YouTube
      • 15. Design Google Drive
      • Loadbalancer Algorithms
      • Cache tier
      • CDN, Content Delivery Network
      • Stateless Web tier
    • Computer System A programmer's perspective
    • Effective Java
      • Item 1. Consider Static Factory Methods Instead of Constructors
      • Item 2. Consider a Builder When Faced with Many Constructor Parameters
      • Item 3. Ensure Singleton with Private Constructor or Enum Type
      • Item 4. Enforce Noninstantiability with a Private Constructor
      • Item 5. Prefer Dependency Injection to Hardwiring Resources
      • Item 6. Avoid Creating Unnecessary Objects
      • Item 7. Eliminate Obsolete Object References
      • Item 8. Avoid Finalizers and Cleaners
      • Item 9.Prefer try-with-resources to try-finally
      • Item10. Adhering to General Rules When Overriding equals
        • Handling Transitivity Issues
        • Ensuring Consistency
      • Item11. Override hashCode When You Override equals
      • Item12. Always Override toString
        • Always Override toString
      • Item13. Override Clone Judiciously
      • Item14. Consider Implementing Comparable
      • Item15. Minimize the Accessibility of Classes and Members
      • Item16. Accessor Methods Over Public Fields
      • Item17. Minimize Mutability
      • Item18. Composition over inherentance
      • Item19. Design and Document for Inheritance, or Else Prohibit It
      • Item20. Prefer Interfaces to Abstract Classes
      • Item21. Design Interfaces with Implementations in Mind
      • Item22. Use Interfaces Only to define Types
      • Item23. Prefer Class Hierarchies to Tagged Classes
      • Item24. Favor Static Member Classes Over Non-Static
      • Item28. Use Lists Instead of Arrays
      • Item29. Prefer Generic Types
      • Item30. Favor Generic Methods
    • Head First Design Patterns
      • Ch1. Strategy Pattern
      • Ch2. Observer Pattern
        • Ver1. Ch2. Observer Pattern
      • Ch3. Decorator Pattern
        • Ch3. Decorator Pattern
      • Ch4. Factory Pattern
      • Ch5. Singleton Pattern
      • Ch6. Command Pattern
      • Ch7. Adapter and Facade Pattern
      • Ch8. Template Method Pattern
    • Digging Deep into JVM
      • Chapter 2. Java Memory Area & Memory Overflow
      • Chapter 3. Garbage Collector & Memory Allocation Strategy (1/2)
      • Chapter 3. Garbage Collector & Memory Allocation Strategy (2/2)
      • Chapter 5. Optimization Practice
      • Chapter 6. Class file structure
      • Chapter 8. Bytecode Executor Engine (1/2)
  • Interview Practices
    • Restful API Practices
      • Url Shortener API
      • Event Ticket Reservation API
      • Course Management API
      • Search posts by tags API
      • Online Code platform API
      • Simple Task Management API
      • Event Participation API
      • Review System API
      • Car management API
      • Online Library
    • Tech Review
      • if(kakao)
        • Kakao Account Cache Migration / if(kakao)2022
        • Improving the Anomaly Detection System for KakaoTalk Messaging Metrics / if(kakao) 2022
        • Standardizing API Case Handling Without Redeployment / if(kakaoAI)2024
        • JVM warm up / if(kakao)2022
    • Naver Computer Science
      • Process & Thread
      • TCP & UDP
      • Spring & Servlet
      • Filter & Interceptor & AOP
      • Equals() & ==
      • Dependency Injection
      • Object Oriented Programming
  • F-Lab
    • Week1
      • Client & Server
      • HTTP
      • TCP/UDP
      • REST API
      • Questions
        • Object Oriented Programming
        • HTTP
        • Process & Thread
        • Data Structure
    • Week2
      • OSI 7 layer
      • Web vs WAS
    • Week3
      • RDB vs NoSQL
      • RDB Index
      • Cache
      • Redis
      • Messaging Queue
    • Week4
      • Project - Ecommerce
    • Week5
      • ERD - 1
    • Week6
      • Ecommerce - 2
      • Role
      • pw hashing && Salt
      • CreatedAt, ModifiedAt
      • JWT
      • Copy of ERD - 1
    • Week7
      • Vault (HashiCorp Vault)
    • Week 8
      • Api Endpoints
    • Week10
      • Product Create Workflow
  • TOY Project
    • CodeMentor
      • Implementation of Kafka
      • Project Improvement (Architectural Enhancements)
      • Communication between servers in msa
  • JAVA
    • MESI protocol in CAS
    • CAS (Compare and Set)
    • BlockingQueue
    • Producer & Consumer
    • Synchronized && ReentrantLock
    • Memory Visibility
    • Checked vs Unchecked Exception
    • Thread
    • Batch delete instead of Cascade
    • Java Questions
      • Week 1(1/2) - Basic Java
      • Week 1(2/2) - OOP
      • Week 2(1/2) - String, Exception, Generic
      • Week2(2/2) Lambda, Stream, Annotation, Reflection
      • Week3(1/2) Collections
      • Week3(2/2) Threads
      • Week4 Java Concurrency Programming
      • Week5 JVM & GC
    • Java 101
      • JVM Structure
      • Java Compiles and Execution Method
      • Override, Overload
      • Interface vs Abstract Class
      • Primitive vs Object Type
      • Identity and equality
      • String, StringBuilder, StringBuffer
      • Checked Exceptions and Unchecked Exceptions
      • Java 8 methods
      • Try-with-reources
      • Strong Coupling and Loose Coupling
      • Serialization and Deserialization
      • Concurrency Programming in Java
      • Mutable vs Immutable
      • JDK vs JRE
  • SPRING
    • DIP. Dependency Inversion Principal
    • Ioc container, di practice
    • @Transactional
    • Proxy Pattern
    • Strategy Pattern
    • Template Method Pattern
    • using profile name as variable
    • Spring Questions
      • Spring Framework
      • Spring MVC & Web Request
      • AOP (Aspect-Oriented Programming)
      • Spring Boot
      • ORM & Data Access
      • Security
      • ETC
  • DATABASE
    • Enhancing Query Performance & Stability - User list
    • Ensuring Data Consistency, Atomicity and UX Optimization (feat.Firebase)
    • Redis: Remote Dictionary Server
    • Database Questions
      • Week1 DBMS, RDBMS basics
      • Week2 SQL
      • Week3 Index
      • Week4 Anomaly, Functional Dependency, Normalization
      • Week5 DB Transaction, Recovery
    • Normalization
      • 1st Normal Form
      • 2nd Normal Form
      • 3rd Normal Form
  • NETWORK
    • HTTP & TCP head of line blocking
    • HTTP 0.9-3.0
    • Blocking, NonBlocking and Sync, Async
    • Network Questions
      • Week1 Computer Network Basic
      • Week2(1/3) Application Layer Protocol - HTTP
      • Week2(2/3) Application Layer Protocol - HTTPS
      • Week2(3/3) Application Layer Protocol - DNS
      • Week3 Application Layer
      • Week4 Transport Layer - UDP, TCP
      • Week5 Network Layer - IP Protocol
    • Network 101
      • https://www.google.com
      • TCP vs UDP
      • Http vs Https
      • TLS Handshake 1.2
      • HTTP Method
      • CORS & SOP
      • Web Server Software
  • OS
    • Operating System Questions
      • Week1 OS & How Computer Systems Work
      • Week2(1/2) Process
      • Week2(2/2) Thread
      • Week3 CPU Scheduling
      • Week4 Process Synchronize
      • Week5 Virtual Memory
    • Operating System 101
      • Operating system
        • The role of the operating system
        • The composition of the operating system.
      • Process
        • In Linux, are all processes except the initial process child processes?
        • Zombie process, orphan process
        • (Linux) Daemon process
        • Process address space
        • Where are uninitialized variables stored?
        • Determination of the size of the Stack and Heap
        • Access speed of Stack vs Heap
        • Reason for memory space partitioning
        • Process of compiling a process
        • sudo kill -9 $CURRENT_PID
      • Thread
        • Composition of a thread's address space
      • Process vs Thread
        • Creation of processes and threads in Linux
      • Multiprocessing
        • Web Browser
        • Implementation of multiprocessing
        • Application areas of multiprocessing
      • Multithreading
        • Application areas of multithreading
      • Interrupt
        • HW / SW Interrupt
        • Method of handling interrupts
        • Occurrence of two or more interrupts simultaneously
      • Polling
      • Dual Mode
        • Reason for distinguishing between user mode and kernel mode
      • System call
        • Differentiation between system calls
        • Types of system calls
        • Execution process of a system call
      • Process Control Block (PCB)
        • PCB의 구조
        • 쓰레드는 PCB를 갖고 있을까?
        • 프로세스 메모리 구조
      • Context switching
        • Timing of context switching
        • Registers saved during context switching
        • Context switching in processes
        • Context switching in threads
        • Difference between context switching in processes and threads
        • Information of the current process during context switching
      • Interprocess Communication (IPC)
        • Cases where IPC is used
        • Process address space in IPC Shared Memory technique
        • Types of IPC
  • COMPUTER SCIENCE
    • Computer Architecture 101
      • 3 components of a computer
      • RAM vs ROM
      • CPU vs GPU
      • SIMD
      • Two's complement
      • Harvard Architecture vs. von Neumann Architecture
      • The structure of a CPU.
      • Instruction cycle (CPU operation method)
      • Instruction pipelining
      • Bus
      • Memory area
      • Memory hierarchy structure
        • Reason for using memory hierarchy structure
      • Cache memory
      • L1, L2, L3 Cache
      • Locality of reference (cache)
      • Fixed-point vs Floating-point
        • epresentation of infinity and NaN (Not a Number) in floating-point
      • RISC vs CISC
      • Hamming code
      • Compiler
      • Linking
      • Compiler vs Interpreter
      • Mutex vs Semaphore
      • 32bit CPU and 64bit CPU
      • Local vs Static Variable
      • Page
  • Programming Paradigm
    • Declarative vs Imperative
  • JPA, QueryDsl
    • why fetchResults() is deprecated
  • PYTHON
    • Icecream
  • FASTAPI
    • Template Page
  • LINUX
    • Template Page
  • DATA STRUCTURE
    • Counting Sort
    • Array vs Linked List
  • GIT, Github
    • git clone, invalid path error
  • INFRA
    • Template Page
  • AWS
    • Server Log Archive Pipeline
    • Image Processing using Lambda
  • DOCKER
    • Docker and VM
    • Python Executable Environment
    • Docker commands
  • docker-compose
    • Kafka, Multi Broker
  • KUBERNATES
    • !Encountered Errors
      • my-sql restarts
      • kafka producer: disconnected
    • Kubernetes Components
    • Helm
      • Helm commands
    • Pod network
    • Service network
      • deployment.yaml
      • services.yaml
    • Service type
      • Cluster IP
      • NodePort
    • service-name-headless?
    • kube-proxy
  • GraphQL
    • Template Page
  • WEB
    • Template Page
  • Reviews
    • Graphic Intern Review
    • Kakao Brain Pathfinder Review
    • JSCODE 자바 1기 Review
  • 😁Dev Jokes
    • Image
      • Plot twist
      • Priorities
      • SQL join guide
      • Google is generous
      • Genie dislikes cloud
      • buggy bugs
      • last day of unpaid internship
      • what if clients know how to inspect
      • its just game
      • how i wrote my achievement on resume
      • self explanatory
      • chr(sum(range(ord(min(str(not))))))
Powered by GitBook
On this page
  1. Books
  2. CS Note for Interview
  3. Ch4. Database

4.7 Principles of Joins

조인의 원리

조인의 종류를 이해한 후, 조인이 데이터베이스에서 어떻게 구현되는지를 이해하는 것도 중요합니다. 데이터베이스는 조인을 수행할 때 다양한 알고리즘을 사용하여 효율적으로 데이터를 결합합니다. 이 섹션에서는 주요 조인 알고리즘인 중첩 루프 조인, 정렬 병합 조인, 해시 조인에 대해 다룹니다.

4.7.1 중첩 루프 조인 (Nested Loop Join)

**중첩 루프 조인(Nested Loop Join)**은 가장 단순한 형태의 조인 알고리즘입니다. 이 방식은 두 테이블을 중첩된 반복문을 통해 결합합니다. 하나의 테이블의 각 행에 대해 다른 테이블의 모든 행을 반복적으로 비교하여 조인 조건을 만족하는 행을 찾습니다. 중첩 루프 조인은 작은 테이블을 조인할 때는 효과적이지만, 대규모 테이블에서는 비효율적일 수 있습니다.

예시: 두 테이블 A와 B가 있다고 가정합니다. 중첩 루프 조인은 다음과 같은 방식으로 작동합니다:

FOR each row in Table A
    FOR each row in Table B
        IF A.column = B.column
            THEN output the row

즉, Table A의 각 행을 가져와 Table B의 모든 행과 비교한 후, 조인 조건이 일치하는 경우 해당 행을 결과에 포함합니다.

SELECT * FROM customers, orders
WHERE customers.id = orders.customer_id;

이 쿼리는 고객 테이블의 각 행에 대해 주문 테이블의 모든 행을 비교하여 조인 조건을 만족하는 행을 찾는 중첩 루프 조인을 사용합니다.

4.7.2 정렬 병합 조인 (Sort-Merge Join)

**정렬 병합 조인(Sort-Merge Join)**은 두 테이블이 조인 조건에 따라 미리 정렬되어 있을 때, 매우 효율적인 조인 알고리즘입니다. 이 방법은 두 테이블의 데이터를 조인 키에 따라 정렬한 후, 각 테이블의 첫 번째 행부터 시작해 비교하며 병합합니다. 정렬 병합 조인은 대규모 데이터 집합을 처리할 때도 효율적입니다.

예시: 두 테이블 A와 B가 각각 정렬되어 있다고 가정합니다. 정렬 병합 조인은 다음과 같이 작동합니다:

1. 두 테이블을 조인 키로 정렬한다.
2. 각 테이블에서 첫 번째 행을 비교한다.
3. 조인 조건이 일치하면 행을 결합하고 다음 행으로 이동한다.
4. 조건이 일치하지 않으면 작은 쪽의 행을 다음으로 이동한다.
5. 테이블의 끝에 도달할 때까지 반복한다.
SELECT * FROM customers
JOIN orders ON customers.id = orders.customer_id
ORDER BY customers.id;

이 쿼리는 두 테이블을 고객 ID로 정렬한 후 병합하여 조인합니다. 이 방식은 특히 두 테이블이 미리 정렬되어 있을 때 매우 효과적입니다.

4.7.3 해시 조인 (Hash Join)

**해시 조인(Hash Join)**은 해시 테이블을 사용하여 조인을 수행하는 알고리즘으로, 조인 조건이 =(등가 조건)일 때 주로 사용됩니다. 이 방법은 두 단계로 나뉩니다: 빌드 단계와 프로브 단계.

  1. 빌드 단계: 작은 테이블의 조인 키를 기반으로 해시 테이블을 생성합니다.

  2. 프로브 단계: 큰 테이블의 각 행에 대해 해시 테이블을 조회하여 일치하는 행을 찾습니다.

예시: 두 테이블 A와 B가 있다고 가정합니다. 해시 조인은 다음과 같이 작동합니다:

1. Table A의 조인 키를 사용하여 해시 테이블을 생성한다. (빌드 단계)
2. Table B의 각 행을 읽어, 해시 테이블에서 일치하는 키를 찾는다. (프로브 단계)
3. 일치하는 키가 있으면 두 행을 결합한다.
SELECT * FROM customers
JOIN orders ON customers.id = orders.customer_id;

이 쿼리는 고객 테이블을 기반으로 해시 테이블을 생성하고, 주문 테이블에서 고객 ID를 사용해 해시 테이블에서 일치하는 값을 찾아 결합합니다. 해시 조인은 특히 테이블이 메모리에 충분히 맞을 때 매우 빠르게 작동합니다.

Previous4.6 Types of JoinsNextCh5. Data Structure

Last updated 9 months ago