antonlnd / cssaturday1 Goto Github PK

The goal here is to build a JavaScript class called FQL that will be a pure JavaScript implementation of a (minimal) SQL-esque database management system and query language. By working on this exercise, you will get a chance to explore how the inner workings of a database engine can work.

Fork and clone this repo, then npm install. Run these tests npm test. IMPORTANT: There are three files where you should be writing code: source/table.js, source/plan.js, and source/fql.js.

At the start, all but the first spec is pending, as specified by xit. As you work through the specs, change those xits to its. Work on these exercises and continue switching off each exercise with your pairing partner.

The role of Table instances is to handle the persistence. In this case that means dealing with the file system, because each table will be a folder where each of its rows is stored in a json file. A Table instance also exposes a .read method for FQL queries to utilize.

The role of FQL instances is to build up a multi-faceted query for a particular table and then return an array of the results when executed with .get. Essentially a query object contains information about what to do later. Only during .get will this information be applied to actually running the query. It does so by using its table's .read and .getRowIds methods.

The role of Plan instances is to simplify the role of queries. Each query should have a plan that contains all of the infromation the query should "remember" for later (when it executes). The plan is also responsible for abstracting some of what a query needs to do when executed. For example, it has .withinLimit method that returns whether or not a possible result is still within the plan's stored limit.