0.0.1 • Published 10 years ago

lsdb v0.0.1

Weekly downloads
3
License
MIT
Repository
github
Last release
10 years ago

LSDB

A TypeScript/JavaScript ORM (currently) backed by localStorage.

Install and set up

Browserify/CommonJS

Run:

npm install --save lsdb

Then for TypeScript:

import LSDB = require('node_modules/lsdb/lsdb')

Or JavaScript:

var LSDB = require('lsdb')

RequireJS/AMD

var LSDB = require('node_modules/lsdb/dist/lsdb-amd')

Usage

LSDB provides a very lightweight ORM over localStorage. Start by setting up a table:

var Pictures = new LSDB.Table('pictures');

you can also provide a unique key (used in update/delete operations):

var Pictures = new LSDB.Table('pictures', 'id');

this is set to 'id' by default. Now you can:

Insert

var pic = {id: 1, caption: "I'm a picture", url: "http://example.com/pic.jpg"};
Pictures.insert(pic);

Retrieve

Pictures.findWhere({id: 1});
// {id: 1, caption: "I'm a picture", url: "http://example.com/pic.jpg"}

Pictures.where({id: 1});
// [{id: 1, caption: "I'm a picture", url: "http://example.com/pic.jpg"}]

Pictures.all();
// [{id: 1, caption: "I'm a picture", url: "http://example.com/pic.jpg"}]

Note that findWhere() returns the first matching record - no warnings/errors for non-unique criteria.

Complex queries

There is also the query() method which returns an Underscore.js chain. You can then use Underscore's extensive methods library to perform complex queries. Lets set up a new table and pretend it has data as per the 'schema':

var Accounts = new LSDB.Table('accounts');
// Schema: {id: number, name: string, age: number, type: string, sex: string}

Now lets look for all customers whose name begins with 'G', sort them by age, and group them by sex:

Accounts.query()
.where({type: 'customer'})
.filter(function(account){
    return /^g/i.test(account.name);
})
.sortBy('age')
.groupBy('sex')
.value();

// {
//      'male': [
//          {id: 23, name: 'Gordon', age: 55, type: 'customer', sex: 'male'},
//          {id: 11, name: 'Graham', age: 23, type: 'customer', sex: 'male'}
//      ],
//
//      'female': [
//          {id: 32, name: 'Gladys', age: 44, type: 'customer', sex: 'female'},
//          {id: 78, name: 'Glynis', age: 19, type: 'customer', sex: 'female'}
//      ]
// }

It's pretty much a fully-featured query interface, praise be unto Underscore.js.

Update

Single record

var pic = Picture.findWhere({id: 1});
pic.caption = "I'm a picture of a kitten";
Pictures.update(pic);
Pictures.findWhere({id: 1});
// {id: 1, caption: "I'm a picture of a kitten", url: "http://example.com/pic.jpg"}

Multiple records

Simply loops over the inputs essentially calling update() for each one, but processes all records before persisting.

var pics = [{id: 1, caption: "One", score: 10},
            {id: 2, caption: "Two", score: 20}];

var attrs = [{id: 1, score: 100}, {id: 2, score: 200}];

Pictures.insertBatch(pics);
Pictures.updateBatch(attrs);
Pictures.all();
// [{id: 1, caption: "One", score: 100}, {id: 2, caption: "Two", score: 200}]

Destroy

Single record

var pics = Pictures.all();
// [{id: 1, caption: "I'm a picture of a kitten", url: "http://example.com/pic.jpg"}]
Pictures.destroy(pics[0].id);
Pictures.all();
// []

'id' is the default unique identifier, so this found the first record with an id of 1 and blasted it.

Multiple records

var Accounts = LSDB.Table('accounts');
Accounts.insertBatch([{id:1, type: 'customer'},{id: 2, type: 'customer'}]);
Accounts.destroyWhere({type: 'customer'});
Accounts.all();
// []

Record Identity

Update and destroy operations require a unique id by which to locate records, by default it is 'id', but you can override the default with your own default by specifying it as the second argument to the constructor:

var Accounts = LSDB.Table('accounts', 'email');

or you can override it per query by specifying it as the second argument:

Accounts.update({card_number: 23232, email: 'someone@example.com'}, 'card_number');

in this case it will find the first record with a card number of 23232 and update it's email address.

NOTE: This is increasingly looking like a questionable design decision, especially as uniqueness of incoming records is not (yet) enforced.

Gotchas, caveats, etc

Internal state

Table data is held in an internal variable called _rows. When you instantiate a new table, LSDB pulls existing data in localStorage into _rows. Updates, inserts, and destroy operations work like this:

  1. Modify _rows
  2. Turn _rows into JSON and persist it to localStorage

which means that an LSDB table instance is unaware of changes it doesn't make, and indeed will overwrite those changes on next save.

Object identity

Everything coming into, or going out of, an LSDB table is deep cloned (via JSON.parse(JSON.stringify(obj))). This means you can't fiddle with LSDB's internal state accidentally, and LSDB can't fiddle with yours.

Record identity

Uniqueness is not enforced!

Contributing

You know the drill.

License

MIT.

0.0.1

10 years ago