mysequence v1.0.4

mysequence

A nodejs sequence generator which generate unique sequential numbers as user-assigned ids for your records/documents.

There are four reasons to use mysequence to generate sequences as ids/keys for your records/documents by application itself.

• Independent: Our data model and persistence layer can be independent of any RDB & noSQL DB such as MySQL auto increment, MongoDB ObjectId, Oracle sequence, MS SQL Server identity, and so on. We can move our data to any store no matter what DB types or providers.
• Fastest: It is the fastest id generator in all known solutions including DBMS proprietary features (like Oracle sequence, MySQL auto increment), UUID utilities, and O/R Mapping built-in features (Hibernate hilo, uuid.string and uuid.hex).
• Unique: UUID utils and O/R Mapping libs CANNOT guarantee unique ids in cluster env.
• Proactive: You know your new record id before you create it, so you can use it in frontend layer proactively.

You can deploy it as an id generating service in/with application code together.

Martin Fowler and peaa

This sequence generator inspired from the design pattern of key table in chapter Identity Field in peaa (Patterns of Enterprise Application Architecture) by Martin Fowler, the famous OOP master.

Martin contributes his great works to the community, and I pay my respect to martin and contribute this OSS lib to the community.

Installation

mysequence need redis as sequence store, by default, but you could also customize your own store.

npm install mysequence

Quick Start

    var SequenceGenerator = require('mysequence').SequenceGenerator;
    var SequenceStore = require('mysequence').SequenceStore;
    var redisClient = require('./redis');
    var store = new SequenceStore({
        keyPrefix: 'seq:id:',
        redis: redisClient,
        logger: console
    });

    var generator = new SequenceGenerator();
    generator.useLogger(console); //set logger here, or just use console as logger
    generator.useStore(store);    //set store here, here we use default redis store

    /*
     * set default sequence config in case of client don't config it.
     */
    generator.useDefaults({
        initialCursor: 0, //default start number for a new sequence
        segment: 20000,   //default segment width which a sequence apply once
        prebook: 18000    //default prebook point when a sequence start to apply a segment in advance
    });

    /**
     * put each of sequence config here
     */
    generator.putAll(
        [{
            key: 'Employee', //the sequence's name which is store in redis.
            initialCursor: 0,     //the sequence's initial value to start from
            segment: 100,         //a number width to increase before sequence touch the segment end.
            prebook: 60           //It means when to book segment. the value is normally
                                  //between half segment (50) and near segment end (90).
        },{
            key: 'User',
            //initialValue: 100,  //by default, use 0
            //segment: 1000,      //by default, use 20000
            //prebook: 500        //by default, use 18000
        }]
    );

    /*
     * Invoke #init method to initialize/sync all sequence to store(redis) until callback
     * is invoked with true parameter.
     */
    generator.init(function(result){
        if(!result){
            throw new Error('generator is not ready');
        }
        var so = generator.get('Employee'); //get a ready sequence object.
        console.log(so.next().val); //get and output a new sequence number
        console.log(so.next().val); //get and output a another sequence number
    });

Performance

Test in my macbook pro 13' (i5 8G), it take < one second to generate 1,000,000 unique numbers using 1000 segment width.

License

MIT License. A copy is included with the source.

Contact

• GitHub (henryleu)
• QQ (1347653)
• Email (henryleu@126.com)