Sklad: Promise-based API for IndexedDB

Sklad library makes work with IndexedDB less weird by providing a tiny Promise-based API on top of IndexedDB. If your browser doesn't support promises you can include polyfill for this.

Starting from 4.0.0 Sklad library is working in all major browsers: Chrome, Firefox, IE11, Microsoft Edge, Safari9 and Android browser. Still there are some browser issues for IE11, Microsoft Edge and Safari9 which can't be patched inside library. Read changelog for more info.

If you're using Sklad with a bundler like Webpack or Rollup, you can either import sklad from 'sklad/es2015' or even import as is if your bundler supports jsnext:main. Otherwise UMD code will be used.

Open database (details)

const conn = await sklad.open(dbName, {
    version: 2,
    migration: {
        '1': (database) => {
            // This migration part starts when your code runs first time in the browser.
            // This is a migration from "didn't exist" to "1" database version
            const objStore = database.createObjectStore('users', {autoIncrement: true});
            objStore.createIndex('fb_search', 'facebook_id', {unique: true});
        },
        '2': (database) => {
            // This migration part starts when your database migrates from "1" to "2" version
            const objStore = database.createObjectStore('users_likes', {keyPath: 'date'});
        }
    }
});

Insert one or multiple records (details)

const conn = await sklad.open(dbName, options);

// insert one document into store
const insertedKey = await conn.insert(objStoreName, 'hello world');

// insert data into multiple stores inside one transaction
const insertedKeys = await conn.insert({
    users: [
        {email: '[email protected]', firstname: 'John'},
        {email: '[email protected]', firstname: 'Jack'},
        {email: '[email protected]', firstname: 'Peter'},
    ],
    foo_obj_store: ['truly', 'madly', 'deeply']
});

assert.equal(insertedKeys, {
    users: [id1, id2, id3],
    foo_obj_store: [id4, id5, id6]
});

Upsert one or multiple records (details)

const conn = await sklad.open(dbName, options);

// upsert one document inside store
const upsertedKey = await conn.upsert(objStoreName, {id: 'BMTH', bandMembersCount: 5})

// upsert data in multiple stores inside one transaction
const upsertedKeys = await conn.upsert({
    users: [
        {email: '[email protected]', firstname: 'John'},
        {email: '[email protected]', firstname: 'Jack'},
        {email: '[email protected]', firstname: 'Peter'},
    ],
    foo_obj_store: ['truly', 'madly', 'deeply']
});

assert.equal(insertedKeys, {
    users: [id1, id2, id3],
    foo_obj_store: [id4, id5, id6]
});

Delete one or mutiple records (details)

const conn = await sklad.open(dbName, options);

// delete document from the object store
await conn.delete(objStoreName, 'key');

// delete multiple documents from different object stores inside one transaction
await conn.delete({
    objStoreName1: ['key_1', 'key_2', 'key_3'],
    objStoreName2: ['key1']
});

Clear one or multiple object stores (details)

const conn = await sklad.open(dbName, options);

// clear everything in one object store
await conn.clear(objStoreName);

// clear everything in multiple object stores
await conn.clear([objStoreName1, objStoreName2]);

Get records from the object store(s) (details)

Beware: if you use index or direction options in your request then fast IDBObjectStore.prototype.getAll() API is not used. This is still okay in most cases. More info here.

const conn = await sklad.open(dbName, options);

// get documents from one object store
const resOneStore = await conn.get(objStoreName, {
    index: 'missing_index', // index name, optional
    direction: sklad.ASC_UNIQUE, // one of: ASC, ASC_UNIQUE, DESC, DESC_UNIQUE, optional
    limit: 4, // optional
    offset: 1, // optional
    range: IDBKeyRange.only('some_key') // range, instance of IDBKeyRange, optional
});

assert.equal(resOneStore, {
    [objStoreName]: [
        {key: ..., value: ...},
        {key: ..., value: ...},
        ...
    ]
});

// get documents from multiple stores in one transaction
const resMultipleStores = await conn.get({
    objStoreName1: {},
    objStoreName1: {limit: 1, offset: 1}
});

assert.equal(resMultipleStores, {
    objStoreName1: [{key: ..., value: ...}, ...],
    objStoreName2: [{key: ..., value: ...}, ...]
});

Count objects in the object store(s) (details)

const conn = await sklad.open(dbName, options);

// count documents inside one object store
const total = await conn.count(objStoreName, {
    range: IDBKeyRange.bound(x, y, true, true), // range, instance of IDBKeyRange, optional
    index: 'index_name' // index name, optional
});

// count documents inside multiple object stores
const res = await conn.count({
    objStoreName1: null,
    objStoreName2: {index: 'index_name'}
});

assert.equal(res, {
    objStoreName1: NUMBER_OF_DOCUMENTS_INSIDE_objStoreName1,
    objStoreName2: NUMBER_OF_DOCUMENTS_INSIDE_objStoreName2
});

Close existing database connection (details)

const conn = await sklad.open(dbName, options);
conn.close(); // it's sync

Delete database

await sklad.deleteDatabase(dbName)

Structuring the database

You can specify keyPath and autoIncrement fields in IDBDatabase.prototype.createObjectStore() inside migration code. Key path (keyPath) is just a name of the field inside the objects which you store in the object store. For instance, it can be foo for object {foo: 'bar'}. Key generator (autoIncrement) is just a name for auto incrementing counter which is used as a primary key. Both of them (key path and key generator) can be used as primary keys for the records stored in the object stores.

More info on MDN.

key path used, no key generator (objects with mandatory field)

const objStore = database.createObjectStore('obj_store_title', {keyPath: 'field_name'});

In this case you must store only objects in the object store. You can specify field_name in the stored objects and its value will be used as a primary key for them. If you don't specify it, Sklad library will generate this field's value for you.

SAMPLE USE CASE: a database of users with unique logins and each user can be represented as an object with fields "firstname", "lastname", "login", "phone" etc.

no key path, key generator used (any data, autoincremented primary key)

const objStore = database.createObjectStore('obj_store_title', {autoIncrement: true});

In this case you can store any type of data in the object store. Primary key for the new created record will be generated by the auto incrementing key generator, but you also can specify your own primary key like this:

const data = sklad.keyValue('your_unique_key', value);
database.insert('obj_store_title', data).then(...);

SAMPLE USE CASE: a simple set of data or even hierarchical objects which don't need a special field to be unique.

key path used, key generator used (objects with optional primary key field)

const objStore = database.createObjectStore('obj_store_title', {
    keyPath: 'field_name',
    autoIncrement: true
});

In this case you must store only objects in the object store. You can specify field_name in the stored objects and its value will be used as a primary key for them. If you don't specify it, its value will be an auto incremented value produced by the key generator.

SAMPLE USE CASE: set of hierarchical objects, which don't have a unique field.

no keypath, no key generator (any data, primary key anarchy)

const objStore = database.createObjectStore('obj_store_title');

In this case you can store any type of data in the object store. You can also specify a key to be used as a primary key for the record like this:

var data = sklad.keyValue('your_unique_key', value);
database.insert('obj_store_title', data).then(...);

Otherwise Sklad library will generate this value for you.

SAMPLE USE CASE: non-structured data with a need for your own primary keys.

Examples

Detailed docs contain nice pieces of code using Sklad library.

Tests

Tests are written with Jasmine testing framework and run with Karma runner. You need to have SauceLabs account to run tests in multiple browsers.

Development/release process

• Watcher is started with yarn watch
• Release files are built with yarn release