rkv

Rkv - reliable kv

Rkv - embeddable KV database in Go (golang).

• Based on Riak bitcast format
• Minimalistic design
• Embeddable and self-contained (no C dependencies)
• Contains both direct and goroutine friendly interfaces
• Use rkv.NewSafe("test.kv") if you want to use with goroutines
• Basic KV admin tool is included in /rkv subfolder, build it and install in your bin folder
• Ability to save records with expiration
• Use Rkv for databases under 50K records

Basic usage:

package main

import (
	"log"
	"strconv"

	"github.com/tadvi/rkv"
)

func main() {
	kv, err := rkv.New("test.kv")
	if err != nil {
		log.Fatal("Can not open database file")
	}
	defer kv.Close()

	type Mytype struct {
		Name string
		Pos  int
	}

	total := 10
	for i := 0; i < total; i++ {
		data1 := Mytype{Name: "one", Pos: 1}
		data2 := Mytype{Name: "two", Pos: 2}
		data3 := Mytype{Name: "three", Pos: 3}

		// this record will be saved for 2 days
		key1 := "key1_" + strconv.Itoa(i)
		err = kv.PutForDays(key1, &data1, 2)
		if err != nil {
			log.Fatal(err)
		}

		// this record will be saved for 3 days
		key2 := "key2_" + strconv.Itoa(i)
		err = kv.PutForDays(key2, &data2, 3)
		if err != nil {
			log.Fatal(err)
		}

		// this record will be saved until removed
		key3 := "key3_" + strconv.Itoa(i)
		err = kv.Put(key3, &data3)
		if err != nil {
			log.Fatal(err)
		}
	}
}

Use rkv tool

Use rkv tool to export or import data. Exported data is in JSON format. This makes it easier to move data around since so many other tools talk JSON.

You can also compact database files with rkv tool.

Basic usage:

$ rkv -c test.kv > test.json

This will compact database and export to JSON.

$ rkv test.kv > test.json

This simply exports database.

$ rkv test.kv < test.json

Imports previously exported database.

Use rkvcsv tool

Basic utility to bring data from relational databases into Rkv. Export your data into CSV then load it into Rkv using this tool.

First row of CSV must have field names.

rkvcsv tool imports CSV files. If all the column's values are numbers then that column is stored as number inside KV. Otherwise it is a string.

Since Rkv is KV datastore you should specify key field that will be used to find record using Get or similar methods. If you do not specify what field in CSV should become record key field inside Rkv then first field is used as key.

Basic usage:

$ rkvcsv -key 0 test.kv < test.csv

Import CSV file and take first field to be a key inside KV database.

$ rkvcsv -key 2 test.kv < test.csv

Import CSV file and take third field from CSV to be a key inside KV database.

Keys and values

Keys are strings. You can use numbers too by simply converting them into strings before you do Put.

Internally values are stored as JSON. So you can store your structs.

Fetch number of keys at once by using GetKeys(with, limit) and then iterate over them and use Get to get actual values like this:

// open database
kv, err := rkv.New("test.kv")
if err != nil {
    log.Fatal("Can not open database file")
}
defer kv.Close()

// get all keys from database - only do this on very small databases!
arr := kv.GetKeys("", -1)	

for _, key := range arr {
    v := Mytype{}
    err := kv.Get(key, &v)
    if err != nil {
        log.Fatal("Error while iterating %q", err.Error())
    }
}

TODO

• Change to use RWMutex instead of Mutex
• Add better iteration techniques
• Add more test cases