Skip to main content Link Menu Expand (external link) Document Search Copy Copied

Data Type of ALG Language

Table of contents

Why Need Data Type

Languages Python and Matlab/Octave are weakly typed which can be convenient for writing the code. However, that is problematic for implementation. The efficiency is not satisfactory compared to C++, and sometimes you may encounter ambiguous error information in Matlab. Therefore, for the sake of efficiency and generality, ALG language is designed to be strongly typed.


The type specification is very simple, because ALG language concentrates on matrices. Basically, the structure of ALG language is prefix + dimension + suffix. For example, f2c means a matrix (dimension is 2) with data type as float and property as a constant.

To specify a variable, add :: between the variable and data type. For example, a::f0 has type f0 (a float scalar).



Basic Type Prefix

Basic type just names the element type.

Prefix Type C++ Type Python Type Matlab/Octave Type
c Complex cx_double complex complex
f Float double double double
i Integer int int int64
u Unsigned Integer uword uint uint64
b Boolean bool bool logical
s String std::string str string
h Character char char char

Alias Prefix

Alias prefixes not only set the element type, but also the dimension. They are the one character alias for a two-character type.

Alias Prefix Type Equivalent Two-character Type
v (Column) Vector c1
r Row Vector c2
m Matrix c2
t Tensor c3
d Double f0

v, r, m and t are all for complex types. For a non-complex type, you need to use the normal two-character way.

Row vector (r) is actually regarded as a matrix for simplicity, so its dimension is still 2. Only column vector (c) is the real vector. But there can be differences in terms of INIT, so it should not be confused with m.


Dimensions range from 0 to 3.

Dimension Type C++ Type
0 Scalar
1 Vector Col
2 Matrix Mat
3 Tensor Cube

Dimension for a scalar can not be omitted.

Please note that matrices are stored in column major order, which is the default order in C++ (Armadillo) and Matlab/Octave. In Python (NumPy), it is equivalent to the option order='F'.

You should always remember the column major order, especially if you use are accustomed to Python. The order will make a big difference to matrix reshape and vectorization.


Suffix Meaning C++ Python Matlab/Octave
c Constant const (None) persistent
r Reference reference (None) (None)

Two suffixes cannot be used together and there is also no need to do so. The use of r is mainly in FUNCTION, allowing a parameter to be changed inside the function body.