Your First Array

Time to create a real array: Rohit Sharma's scores from his first 5 matches. Everything you will do in NumPy starts with an array like this.

Definition

np.array(object) is the function that creates an ndarray from existing Python data — usually a list (or a list of lists). It copies the values into a new contiguous block of memory and automatically picks a single dtype that fits all of them.

• object — the data to convert: a list, tuple, nested list, or another array.
• copies — changing the array later does not change the original list (and vice versa).
• automatically picks a dtype — whole numbers → int64, any decimal → float64, True/False → bool.

Syntax

np.array(object, dtype=None)

(Full details of np.array are in the Creating Arrays chapter.)

Example — create and inspect

import numpy as np

rohit_runs = np.array([56, 0, 74, 88, 32])
print(rohit_runs)
print(type(rohit_runs))
print(rohit_runs.dtype)
print(rohit_runs.shape)
print(repr(rohit_runs))

[56  0 74 88 32]
<class 'numpy.ndarray'>
int64
(5,)
array([56,  0, 74, 88, 32])

How it works:

• print(arr) shows values separated by spaces, not commas — that is how you recognise a NumPy array in printed output. Numbers are right-aligned ( 0) so columns line up.
• type(...) confirms it is a numpy.ndarray.
• .dtype is int64 — 64-bit integers, because all values are whole numbers.
• .shape is (5,) — a 1-D array with 5 elements. The trailing comma means it is a one-element tuple.
• repr(arr) (what a Jupyter notebook shows when you type just rohit_runs) prints array([...]) with commas.

Example — use it: index, slice, maths, change values

import numpy as np
rohit_runs = np.array([56, 0, 74, 88, 32])
print(rohit_runs[0])
print(rohit_runs[-1])
print(rohit_runs[1:4])
print(rohit_runs + 5)
print(rohit_runs.sum(), rohit_runs.max(), rohit_runs.min())
print(len(rohit_runs))
rohit_runs[1] = 12
print(rohit_runs)
rohit_runs[2] = 99.7
print(rohit_runs)
print(rohit_runs.tolist())

56
32
[ 0 74 88]
[61  5 79 93 37]
250 88 0
5
[56 12 74 88 32]
[56 12 99 88 32]
[56, 12, 99, 88, 32]

How it works:

• Indexing ([0], [-1]) and slicing ([1:4], stop excluded) work exactly like Python lists.
• rohit_runs + 5 adds 5 to every score and returns a new array; the original is unchanged.
• .sum(), .max(), .min() are methods built into every array. len() still works and gives the length of the first dimension.
• Arrays are mutable: rohit_runs[1] = 12 changes the value in place.
• Assigning 99.7 into an integer array silently truncates it to 99 — the dtype never changes on assignment.
• .tolist() converts back to a normal Python list (with commas).

Printed output: how to tell what you have

Common mistakes

• ❌ np.array(56, 0, 74) — passing values as separate arguments. ✅ Wrap them in a list: np.array([56, 0, 74]).
• ❌ Expecting arr[2] = 99.7 to store a decimal in an int array. ✅ Create the array as float (np.array([...], dtype=float)) if you need decimals.
• ❌ Forgetting that arr + 5 returns a new array. ✅ Store it: arr = arr + 5 (or arr += 5) if you want to keep the result.

Interview tip

Q: Why does np.array([1, 2, 3])[0] = 9.99 store 9? A: An array's dtype is fixed at creation. Assigning a float into an int64 array casts the value to int, which truncates toward zero. To keep decimals, create the array with a float dtype.