Learn Functional Programming by Implementing SQL with Underscore.js

When web developers first approach functional programming libraries, like underscore.js, they are presented a scattered set of pieces without guidance on how they might fit together to form a complete and beautiful picture. As it turns out, most web developers are familiar with a “big picture” composition of functional programming pieces in the form of SQL.

The basic execution of a SQL query can be decomposed into a series of steps that join, filter, order, group, and transform collections of data. We will breaking down a SQL query into a pipeline of steps. At each step we’ll write an implementation in JavaScript using basic, functional techniques leveraging a breadth of the underscore.js library along the way. As we move along we will recompose our JavaScript pieces to fit beautifully together.

By the conclusion of the talk we will have implemented a basic SQL-like DSL for querying collections of in-memory objects in JavaScript. By translating your knowledge of SQL into functional programming terms you will more easily identify practical uses and compositions of functional programming concepts. You will also approach SQL from a fresh perspective of an implementer and viscerally learn things like why web scale databases avoid JOINs like the plague and why that one query of yours is might be running so slow.

Outline:

1. Why care about functional programming in JavaScript?
   1. Small, simple, testable pieces
   2. Powerful, practical, flexible compositions
2. How does a SQL query work?
   1. `from` and `join` identify the sources
   2. `where` filters data based on criteria
   3. `order by` sorts data based on values and direction
   4. `select` projects a subset of the data
3. Translating our SQL knowledge into Functional Programming techniques
   1. Brief overview of higher-order JavaScript functions and function generators
   2. Applying underscore.js’ `filter` function to write `where`
   3. Wraping around `sortBy`
   4. Rolling our own `project` with `each` and `map`
4. Composing our pieces into a simple DSL
   1. Bringing it all together
   2. So why are JOINs so scary for web scale databases?
5. Conclusion and Questions