🚂 Try elena

go mod tidy
go run ./cmd/elenadb mydb.elena

elena>

ElenaDB internal workflow (spanish)

The modules

• The Elena CLI:

  See docs/cli.md for the Elena CLI specification.

  go run ./cmd/elena mydb.elena

  The REPL communicates directly with the Execution Engine, starting from the parsing stage. Currently only parsing is supported.

• The Elena Query Language:

  See docs/queries.md for the complete Language specification.

  • SQL-like syntax heavily inspired by the peruvian lexicon.

  • Easy to understand and know what's happening.

  • Compound queries were never so easy to read!

    let pedro = dame { id } de usuario donde (nombre=="pedro") pe
    
    mete {
        id_user: pedro.id,
        document_type: 'DNI',
        document_number: '72016572',
    } en doctor pe
    

• Storage Engine

  We'll build a disk-oriented storage manager for the Elena DBMS. The primary storage location of the database is on disk. The BufferPool in the storage manager allows the DBMS to handle databases larger than the available system memory by transparently managing page requests.

  Key Components:

  1. LRU-K Replacement Policy Objective: Track page usage and implement an eviction policy based on backward k-distance.
  2. Disk Scheduler Objective: Schedule read and write operations on the disk. Make sure it is thread-safe!
  3. Buffer Pool Manager Objective: Manage fetching and storing of database pages between disk and memory.

Algorithms used

• LRU-K eviction policies: A page replacement policy meant to solve the problems that LRU has, such as sequential flooding. https://en.wikipedia.org/wiki/Page_replacement_algorithm#Variants_on_LRU

• Disk scheduling: Used to optimize and prioritize disk accesss. https://en.wikipedia.org/wiki/I/O_scheduling

• Tree traversing: BFS and DFS are used to traverse the B+Tree. https://en.wikipedia.org/wiki/Tree_traversal

• Greedy optimization: By generating all the possible plans to fulfill a query, we can choose the best one. https://en.wikipedia.org/wiki/Greedy_algorithm

• Heap Sort: Used to sort the results of a query as they are being iterated through the query plan. https://en.wikipedia.org/wiki/Heapsort

Data structures used

• Bloom filters. A probabilistic data structure to test whether an element is on a set. Used to implement fast Hash Table misses. https://en.wikipedia.org/wiki/Bloom_filter

• Hash Tables: - [] Used in Hash Joins operations. - [X] Used in Page Table, in Buffer Pool Manager.

• Concurrent queue: Thread safe queue implementation.

• Finite State Machine (FSM)