MATLAB Variables and Data Types

Analogy categoryAnalogies

Variables in MATLAB are created simply by assignment — there's no separate declaration step and no explicit type keyword, since MATLAB is dynamically typed and infers the type from the value on the right-hand side of '='; if you type an expression without assigning it to a name, MATLAB automatically stores the result in a special variable called ans.

🏏

Cricket analogy: Not needing to declare a variable's type before assigning it is like a scorer jotting down a batter's runs the moment the ball is hit, without first filling out a form declaring what kind of shot it will be — the type, four, six, or single, is inferred from what actually happens.

Numeric Types: double, single, and Integers

MATLAB's default numeric type is double (a 64-bit IEEE 754 floating-point number) — even typing x = 5 creates a double, not an integer — but you can explicitly create other numeric types like single, int8, uint8, int16, int32, and int64 by wrapping a value in the corresponding function, e.g. int32(5), which is important when working with image data (typically uint8, 0-255) or when you need to conserve memory on very large arrays.

🏏

Cricket analogy: Using double by default and only switching to int8 when needed is like a team defaulting to their standard Test-match XI but switching to a specialist night-watchman only when the situation specifically calls for it.

matlab

% Dynamic typing: type is inferred from the assigned value
x = 5;              % double by default
class(x)             % 'double'

% Explicit numeric types
px = uint8(200);      % typical for image pixel data (0-255)
bigCount = int64(9000000000);

% Text types
name1 = 'Ada';         % char array
name2 = "Ada";         % string object
isequal(name1, name2)   % true, values compare equal despite different types

% Logical type from comparison
isPassing = (75 >= 60)  % logical: 1 (true)

% Struct: a single record with named fields
student.name = 'Priya';
student.grades = [88 92 79];

% Cell array: mixed types and sizes
mixedData = {'Priya', [88 92 79], true};

Run whos in the Command Window (no arguments) to see every variable currently in your Workspace along with its size in bytes, dimensions, and class — this is the fastest way to spot an accidental type mismatch, like a double where you expected a logical.

Text, Logical, and Character Data

MATLAB has two text types: the older char array (single-quoted, e.g. 'hello', stored as a row vector of character codes) and the newer string type (double-quoted, e.g. "hello", introduced in R2016b, which supports array-of-strings semantics and functions like strsplit and contains more naturally); the logical type stores only true/false values and is what comparison operators like == and < return, and is also used for logical indexing.

🏏

Cricket analogy: The distinction between char and string types is like the difference between a scorecard written as loose individual chalk marks, a row of character codes, versus a proper digital scoring app that treats 'India 287/6' as one complete entry you can search and split.

Container Types: Cell Arrays and Structures

For data that isn't uniform in size or type, MATLAB provides the cell array (created with curly braces {}), which can hold different types and sizes in each element — a mix of strings, matrices, and numbers — and the struct, which groups related data under named fields accessed with dot notation (e.g. student.name, student.grades), making it the natural choice for representing a single record with several attributes.

🏏

Cricket analogy: A cell array is like a kit bag that can hold a bat, gloves, and a water bottle side by side even though they're completely different shapes and sizes, while a struct is like a player's official profile card with named fields, name, team, battingAverage, each holding a different kind of data.

Comparing text with == behaves differently for char and string: 'cat' == 'cat' returns a logical array (one comparison per character), while "cat" == "cat" returns a single true — mixing the two habits is a common source of subtle bugs, so prefer strcmp for char arrays and == only for string objects.

MATLAB is dynamically typed -- no declaration keyword is needed, and type is inferred from the assigned value.
double is the default numeric type; narrower types like int8/uint8 must be created explicitly.
char arrays (single-quoted) and string objects (double-quoted) are distinct text types with different comparison behavior.
logical is the type returned by comparison operators and used for logical indexing.
Cell arrays ({}) hold mixed types and sizes; structs group named fields into one record via dot notation.
whos lists every variable's size, dimensions, and class in the current Workspace.
Use strcmp for char array comparisons; == works directly and safely only on string objects.