Elixir Expert
Core Philosophy
- Immutability by default — data transforms through pipelines, never mutates
- Let it crash — design for recovery via supervision, not defensive coding
- Explicit over implicit — no magic, no hidden state
- Composition over inheritance — behaviours + protocols, not class hierarchies
- Small, pure functions composed via pipes
Decision Tree: Data Structures
Need key-value data?
├── Fixed keys known at compile time? → Struct (enforce shape)
├── Dynamic/unknown keys? → Map
├── Ordered options/small config? → Keyword list
├── Need default values? → Keyword list with Keyword.get/3
└── Passing to Ecto/Phoenix? → Check what the API expects
Decision Tree: Polymorphism
Need different behavior per data type?
├── Dispatching on data TYPE (struct)? → Protocol
│ ├── You own the types? → defimpl in each module
│ └── Third-party types? → defimpl in your protocol module
├── Dispatching on MODULE (callback contract)? → Behaviour
│ ├── Swappable implementations (prod/test)? → Behaviour + Mox
│ └── Plugin system? → Behaviour + Registry
└── Simple branching on values? → Pattern matching (function heads/case)
Decision Tree: Control Flow
Which construct?
├── Multiple sequential operations that can fail? → with
├── Single value, multiple patterns? → case
├── Multiple boolean conditions? → cond
├── Two outcomes? → if/else (only for simple boolean)
├── Type/value dispatch? → Function heads with guards
└── Never: nested if/case — refactor to function heads or with
Decision Tree: Error Handling
How to handle this error?
├── Business logic failure (expected)? → {:ok, val} / {:error, reason}
│ ├── Chain of operations? → with + pattern match on errors
│ └── Single operation? → case on result tuple
├── Programmer error (bug)? → raise / assert
├── External system failure? → {:error, reason} + let caller decide
├── Process crash (let it crash)? → Don't rescue, let supervisor restart
└── Must clean up resources? → try/after (rare)
Pattern Matching Essentials
# Function heads — preferred over case inside functions
def process(%User{role: :admin} = user), do: admin_path(user)
def process(%User{role: :member} = user), do: member_path(user)
def process(_), do: {:error, :invalid_user}
# Pin operator — match against existing binding
expected = "hello"
^expected = get_value() # asserts equality
# Guards — extend pattern matching
def fetch(id) when is_binary(id), do: Repo.get(Thing, id)
def fetch(id) when is_integer(id), do: Repo.get(Thing, Integer.to_string(id))
# Custom guards
defguard is_positive(value) when is_number(value) and value > 0
Data Transformation Idioms
# Pipeline — always flows data left-to-right
data
|> Enum.filter(& &1.active)
|> Enum.map(&transform/1)
|> Enum.sort_by(& &1.name)
# Enum vs Stream
# Enum: eager, for bounded collections (99% of cases)
# Stream: lazy, for large/infinite sequences or multiple passes
# Access — nested data traversal
get_in(data, [:user, :address, :city])
update_in(data, [:user, :name], &String.upcase/1)
Module Organization
defmodule MyApp.Accounts.User do
@moduledoc "User account with authentication and profile data."
# 1. use/import/alias/require (in this order)
use Ecto.Schema
import Ecto.Changeset
alias MyApp.Accounts.Organization
# 2. Module attributes
@primary_key {:id, :binary_id, autogenerate: true}
# 3. Schema / struct / type definitions
schema "users" do
# ...
end
# 4. Public API functions
# 5. Private functions (at bottom)
end
Common Gotchas
- Atoms are not garbage collected — never convert user input to atoms
- Large binaries — sub-binaries reference parent; copy if parent is large
- Timezone — always store UTC, convert at boundaries (DateTime, not NaiveDateTime)
- Decimal — use
Decimal for money, never Float
- String vs charlist —
"hello" (binary) vs 'hello' (charlist/Erlang), prefer binary
References
references/idioms.md — Pipeline patterns, comprehensions, recursive patterns
references/protocols-behaviours.md — Full protocol/behaviour implementations
references/error-handling.md — With chains, error structs, changeset errors