50 lines (40 loc) · 1.48 KB

Functions

Simple function is always good

Once you develop your function, it gets bigger and bigger.
- You must continuously reorganize the function structure
Refer to Databricks Spark Scala Guide link
- <30 lines in a function is the best.

DO ONE THING ; command-query seperation

That is, just only do what the name indicates
if the function is dealing both command and query, so the name should
- e.g. decrementAndReturnIfZero()

Use descriptive names

e.g. obtainSomethingAndElseToReturnThis()

In switch statements

try to use enum types for easy comprehension

Arguments

In scala, try to use named argument when the function has multiple arguments with same type
Good arguments requires a lot of conceptual power
- Order, Type, Options
Output arguments is even harder

# of Arguments

1: useful when in/out type matches ; transformation operation
2: ordering problem, especially when comparing the two?
3: harder!
Reducing # arguments by introducing a class is a good practice

Exception Handling

A must-do practice
Define error codes (object or enum) in your own way

Extract try-catch blocks

def wrapperFunction():
  try:
    innerFunction()
  catch:
    (excpetional flow)

Structured programming

Only 1 return statement
No break/continue statement
Every control flow must follow the structure!

RENAME / SPLIT / CHANGE NAME / SHRINK / REORDER!