Libraries like NumPy
NumPy ndarrays
NumPy’s core data structure is the ndarray (N-dimensional array). It’s a powerful tool for numerical computation because it provides efficient storage and manipulation of large arrays of homogeneous data.
Key features:
-
Homogeneous data type: All elements within a single ndarray must be of the same data type (e.g.,
int32,float64, etc.). This allows for optimized memory usage and vectorized operations. -
N-dimensional: ndarrays can represent data in any number of dimensions (1D vectors, 2D matrices, 3D tensors, and beyond).
-
Vectorized operations: NumPy allows for performing operations on entire arrays at once, rather than iterating element by element. This significantly improves performance.
-
Broadcasting: NumPy’s broadcasting rules allow for operations between arrays of different shapes under certain conditions, simplifying code and improving efficiency. Broadcasting in NumPy
-
Slicing and indexing: Powerful slicing and indexing mechanisms enable accessing and manipulating portions of ndarrays efficiently. NumPy Array Indexing and Slicing
-
Memory efficiency: ndarrays store data in contiguous memory locations, improving performance and reducing memory overhead compared to Python lists.
Example:
import numpy as np
# Creating a 2D array
arr = np.array(1, 2, 3, 4, 5, 6)
# Accessing elements
print(arr[0, 1) # Output: 2
# Performing vectorized operations
print(arr * 2) # Output: [ 2](./../-2/) 4 6, [ 8 10 12]]
# Array shape
print(arr.shape) # Output: (2, 3)
# Array data type
print(arr.dtype) # Output: int64 (or similar, depending on your system)
Related Notes: