litesql
v0.0.3
Published
The easy way to deal with sqlite databases in node.js
Downloads
53
Maintainers
Readme
What's this ?
litesql is a tiny library easing developpement with sqlite databases.
work was based upon the massive-js library. massivejs is a nodejs module and provides an intuitive interface over raw Mysql and Postgre modules.
basically, litesql intends to brings the same interface into sqlite database. undeway the module use the sqlite3 module, meaning you can access all the functions it offers.
Of course you may also use the additional facilities this module offers (otherwise, you wouldn't be here, right?)
Getting started
install it
npm install litesql
then use it
var litesql = require('litesql');
var db = litesql.db(':memory:');
db.serialize(function() {
db.createTable(
// table name
'todos',
// table defintion
{
// shortcut for id INTEGER PRIMARY KEY AUTOINCREMENT
id: 'pk',
num: 'int',
// column definition can be an object too; you can pass it also 'unique: true'
task: { type: 'text', required: true },
// type alias is managed internally by the library
duedate: 'date',
completed: 'boolean'
}
).run();
/*
you can also write
var query = db.createTable(...);
query.run( function (err) { ... } );
*/
// helper class
var todos = new litesql.Table('todos', 'id', db);
for(var i = 1; i <= 10; i++) {
todos.insert({ num: i, task: 'Task #'+i, duedate: new Date(), completed: false }).run();
}
todos.find().all( function(err, tasks){
assert.equal(tasks.length, 10);
});
})
So basically, it works always the way you've seen it
- Construct a query via helper methods and classes
- then execute the query using methods like you are used to ( run, get, all, each )
CRUD methods
We've already seen insert; following how to update an existing record given its primary key; usually (but not necessary) an 'id' column;
db.serialize(function() {
// update by pk (id = 1)
todos.update({ task: 'have to finish this' }, 1 /* pk */).run();
// and also find by pk (id = 1)
todos.find(1).get(function(err, todo) {
assert.equal(todo.task, 'have to finish this');
});
});
You can also update by another condition; here we update all records with num <= 5
db.serialize(function() {
// update all completed todos
// set compteted = true on all records with num <= 5
todos.update({ completed: true }, { 'num <=': 5 }).run();
// we can also call #find with an object hash for conditions
todos.find({ completed: true }).all(function(err, completedTasks) {
assert.equal(completedTasks.length, 5);
});
});
Another way to insert a new record is via the #save method
db.serialize(function() {
// will insert a new record, since there is no pk field
todos.save({ task: 'give me more examples', completed: false }).run();
todos.find({ task: 'give me more examples' }).all(function(err, tasks) {
assert.equal(tasks.length, 1);
});
});
You can use #save to update an existing record as well. Just include the pk field
db.serialize(function() {
// will update an existing record, since we have specified the pk field
todos.save({ id: '1', task: 'first of firsts' }).run();
todos.find(1).get(function(err, todo) {
assert.equal(todo.task, 'first of firsts');
});
});
We use #remove to delete an existing record; below we remove by the pk field
db.serialize(function() {
// remove todo by pk (id=10)
todos.remove(10).run();
todos.find(10).all(function(err, todos) {
assert.equal(todos.length, 0);
});
});
As you may have already guessed, you can call #remove with more conditions; below we remove all tasks with completed=true
db.serialize(function() {
// remove all completed tasks
todos.remove({ completed: true }).run();
todos.find({ completed: true }).all(function(err, todos) {
assert.equal(todos.length, 0);
});
});
Schema helper methods
TBD