Log Structured Database Engine

A toy implementation of a log structured database engine, inspired from Designing Data Intensive Applications (DDIA) book by Martin Kleppmann.

Martin shows that you can create the world's simplest database using the follow lines of bash

db_set() {
    echo "$1,$2" >> database
}

db_get() {
    grep "^$1," database | sed -e "s/^$1,//" | tail -n 1
}

This is my own not-so-concise implementation in Go, used to solidify the concepts that Martin portrays. Comments are provided for explanation and understanding.

Example

Build using go build cmd/db.go and then run

./db --set "1, foo"
./db --set "2, bar"
./db --get "1" # outputs 'foo'
./db --set "1, bar" # updates ID 1 to bar
./db --get "1" # outputs 'bar'
./db --disable-index --get "1" # also outputs 'bar', but with a full scan returning the latest record