Title: NumPy Arithmetic Functions & Broadcasting Notes
1. Using Arithmetic Functions in NumPy
Import NumPy:
import numpy as np
Minimum and Maximum values:
• np.min(array) → minimum value
• np.max(array) → maximum value
Example:
arr = np.array([1, 2, 3, 4, 5])
print(np.min(arr)) # 1
print(np.max(arr)) # 5
Index positions of Min and Max:
• np.argmin(array) → index of minimum
• np.argmax(array) → index of maximum
Multi-dimensional arrays:
• Use axis to get min/max along rows or columns
arr = np.array([[1, 2, 3], [4, 5, 6]])
print(np.min(arr, axis=0)) # column-wise min → [1, 2, 3]
print(np.min(arr, axis=1)) # row-wise min → [1, 4]
Square root:
• np.sqrt(array) → element-wise square root
Trigonometric functions:
• np.sin(array) → sine
• np.cos(array) → cosine
1
x = np.array([0, np.pi/2, np.pi])
print(np.sin(x)) # [0, 1, 0]
print(np.cos(x)) # [1, 0, -1]
Cumulative sum:
• np.cumsum(array) → cumulative sum
arr = np.array([1, 2, 3, 4])
print(np.cumsum(arr)) # [1, 3, 6, 10]
Real-life uses: - Min/Max → find extremes - ArgMin/ArgMax → find positions (e.g., max sales day) -
Cumulative sum → statistics, time series, progress tracking - Trig & sqrt → scientific computations
2. Array Shape in NumPy
Checking shape:
arr = np.array([[1, 2], [3, 4]])
print(arr.shape) # (2, 2) → 2 rows, 2 columns
Converting 1D → 2D → 3D:
arr = np.array([1,2,3,4,5,6])
print(arr.shape) # (6,)
arr2d = arr.reshape(2,3)
print(arr2d.shape) # (2,3)
arr3d = arr.reshape(2,1,3)
print(arr3d.shape) # (2,1,3)
Checking dimensions:
print(arr.ndim) # 1
print(arr2d.ndim) # 2
print(arr3d.ndim) # 3
2
Flatten back to 1D:
flat = arr3d.reshape(-1)
print(flat) # [1 2 3 4 5 6]
print(flat.shape) # (6,)
Rules:
• .shape → gives rows, columns, structure
• .ndim → number of dimensions
• .reshape() → convert arrays (element count must remain same)
• .reshape(-1) → flatten to 1D
3. Broadcasting in NumPy
Definition: - Allows arithmetic operations on arrays of different shapes. - Smaller array is automatically
stretched to match shape of larger array (without copying data).
Broadcasting Rules:
1. Compare dimensions from right to left.
2. Dimensions must be either same or 1.
3. Result shape = maximum size along each dimension.
4. If rules don’t match → ❌ error
Example 1: Simple Addition
import numpy as np
a = np.array([1, 2, 3, 4]) # shape (4,)
b = np.array([10]) # shape (1,)
print(a + b) # [11 12 13 14]
Example 2: 2D + 1D
a = np.array([[1,2,3],[4,5,6]]) # (2,3)
b = np.array([10,20,30]) # (3,)
print(a + b)
# [[11 22 33]
# [14 25 36]]
3
Example 3: Mismatched shapes
a = np.array([[1,2,3],[4,5,6]]) # (2,3)
b = np.array([1,2]) # (2,)
print(a + b) # ❌ Error
Summary: - Broadcasting enables operations on arrays of different shapes. - Compare dimensions right →
left. - Dimensions must be same or 1. - Result shape = maximum along each dimension. - Mismatch → error.
