Advanced Array Creation in NumPy – Complete Tutorial (Hinglish)
Aaj hum Python ki NumPy library mein Advanced Array Creation seekhenge. Main har concept ko basic se explain karunga, with practical examples.
Har code line ke explanation ko usi line ke immediately side mein rakhenge, taaki tumhe code aur uska meaning ek saath samajh aaye.
1. Introduction: Advanced Array Creation kya hai?
NumPy mein hum np.array() se arrays create kar sakte hain. Lekin NumPy humein kuch special functions bhi provide karta hai jinse hum easily different types ke arrays bana sakte hain.
For example:
Numbers ki sequence create karna.
Equal intervals ke saath numbers generate karna.
Zero aur one se bhare arrays banana.
Kisi particular number se array fill karna.
Identity matrix create karna.
2D, 3D aur multidimensional arrays banana.
Array ka data type control karna.
Sabse pehle NumPy import karte hain.
import numpy as np # NumPy library import kar rahe hain aur uska short name np rakh rahe hainAb hum ek-ek function ko detail mein samajhte hain.
2. np.arange()
Concept: np.arange() ka use numbers ki sequence generate karne ke liye hota hai.
Ye Python ke range() function ki tarah kaam karta hai, lekin ye NumPy array return karta hai.
Syntax:
np.arange(start, stop, step)Parameter | Meaning |
|---|---|
start | Sequence kahan se start hogi |
stop | Sequence kahan tak jayegi (excluded) |
step | Har number ke beech ka difference |
Example 1: Basic array creation
import numpy as np # NumPy import kar rahe hain
arr = np.arange(1, 10, 1) # 1 se start, 10 se pehle stop, har baar 1 add hoga
print(arr) # Generated array print karengeOutput:
[1 2 3 4 5 6 7 8 9]Samjho:
1→ Starting number.10→ Ending limit, lekin 10 include nahi hoga.1→ Har next number mein 1 add hoga.
Example 2: Step size change karna
arr = np.arange(2, 20, 2) # 2 se start, 20 se pehle stop, har baar 2 add hoga
print(arr) # Array print karengeOutput:
[ 2 4 6 8 10 12 14 16 18]Yahan har number ke beech difference 2 hai.
Example 3: Negative numbers
arr = np.arange(-5, 6, 1) # -5 se start karke 6 se pehle tak numbers generate honge
print(arr) # Array print karengeOutput:
[-5 -4 -3 -2 -1 0 1 2 3 4 5]Example 4: Decimal step size
arr = np.arange(0, 2, 0.5) # 0 se start, 2 se pehle tak 0.5 ke gap mein values
print(arr) # Array print karengeOutput:
[0. 0.5 1. 1.5]Real-life use: Agar tumhe 1 se 100 tak even numbers generate karne hain, toh np.arange() bahut useful hai.
Important: stop value include nahi hoti.
3. np.linspace()
Concept: np.linspace() ka use ek fixed starting point aur ending point ke beech equally spaced numbers generate karne ke liye hota hai.
Iska main difference np.arange() se ye hai ki yahan hum decide karte hain ki total kitne numbers chahiye.
Syntax:
np.linspace(start, stop, num)Parameter | Meaning |
|---|---|
start | Starting value |
stop | Ending value |
num | Total kitne numbers generate karne hain |
Example 1: Basic usage
arr = np.linspace(1, 10, 5) # 1 se 10 ke beech equally spaced 5 numbers generate honge
print(arr) # Array print karengeOutput:
[ 1. 3.25 5.5 7.75 10. ]Kaise kaam kiya?
Humein 1 se 10 ke beech total 5 numbers chahiye.
NumPy automatically equal spacing calculate karega.
Number distribution samjho
1
3.25
5.5
7.75
10
Example 2: Total 10 numbers generate karna
arr = np.linspace(0, 100, 10) # 0 se 100 ke beech total 10 equally spaced values
print(arr) # Generated array display karengeOutput:
[ 0. 11.11111111 22.22222222 33.33333333
44.44444444 55.55555556 66.66666667 77.77777778
88.88888889 100. ]Example 3: Endpoint exclude karna
linspace() mein endpoint parameter bhi hota hai.
arr = np.linspace(1, 10, 5, endpoint=False) # endpoint=False se ending value 10 include nahi hogi
print(arr) # Array print karengeOutput:
[1. 2.8 4.6 6.4 8.2]np.arange() vs np.linspace()
np.arange() | np.linspace() |
|---|---|
Step size define karte hain | Total numbers define karte hain |
Stop value default se excluded hoti hai | Stop value default se included hoti hai |
Integer sequences ke liye useful | Equal spacing ke liye useful |
|
|
Real-life use: Machine learning, scientific calculations, graphs aur mathematical functions mein linspace() bahut useful hai.
4. np.zeros()
Concept: np.zeros() ka use aisa NumPy array create karne ke liye hota hai jisme saari values zero hoti hain.
Syntax:
np.zeros(shape)Yahan shape decide karta hai ki array ka size aur dimensions kya honge.
Example 1: 1D zero array
arr = np.zeros(5) # 5 elements ka array create hoga, sabhi values zero hongi
print(arr) # Array print karengeOutput:
[0. 0. 0. 0. 0.]Notice karo ki output mein 0. aa raha hai, kyunki default dtype float64 hota hai.
Example 2: 2D zero array
arr = np.zeros((2, 3)) # 2 rows aur 3 columns ka zero array create hoga
print(arr) # 2D array print karengeOutput:
[[0. 0. 0.]
[0. 0. 0.]]Yahan:
2 = Rows
3 = Columns
Example 3: Integer zero array
arr = np.zeros(5, dtype=int) # dtype=int se values integer type mein create hongi
print(arr) # Array print karengeOutput:
[0 0 0 0 0]Real-life use:
Maan lo tum ek student marks system bana rahe ho aur initially sabhi students ke marks zero rakhna chahte ho. Tab np.zeros() useful hai.
5. np.ones()
Concept: np.ones() ka use aisa array create karne ke liye hota hai jisme sabhi elements ki value 1 hoti hai.
Syntax:
np.ones(shape)Example 1: 1D array
arr = np.ones(5) # 5 elements ka array create hoga, sabhi values 1 hongi
print(arr) # Array print karengeOutput:
[1. 1. 1. 1. 1.]Example 2: 2D array
arr = np.ones((3, 2)) # 3 rows aur 2 columns ka array create hoga
print(arr) # Array print karengeOutput:
[[1. 1.]
[1. 1.]
[1. 1.]]Example 3: Integer ones array
arr = np.ones((2, 3), dtype=int) # 2x3 array create hoga aur values integer hongi
print(arr) # Array print karengeOutput:
[[1 1 1]
[1 1 1]]Real-life use: Initialization, mathematical calculations aur machine learning mein initial values assign karne ke liye useful hai.
6. np.full()
Concept: np.full() ka use aisa array create karne ke liye hota hai jisme hum apni choice ki koi bhi value fill kar sakte hain.
zeros() mein sabhi values 0 hoti hain.
ones() mein sabhi values 1 hoti hain.
Lekin full() mein hum koi bhi value de sakte hain.
Syntax:
np.full(shape, fill_value)Example 1: Value 5 se array fill karna
arr = np.full(5, 5) # 5 elements ka array create hoga aur har element ki value 5 hogi
print(arr) # Array print karengeOutput:
[5 5 5 5 5]Example 2: 2D array with value 10
arr = np.full((2, 3), 10) # 2 rows aur 3 columns ka array banega, sabhi values 10 hongi
print(arr) # Array print karengeOutput:
[[10 10 10]
[10 10 10]]Example 3: Decimal values
arr = np.full((2, 2), 5.5) # 2x2 array create hoga aur har element mein 5.5 hoga
print(arr) # Array print karengeOutput:
[[5.5 5.5]
[5.5 5.5]]zeros(), ones() aur full() mein difference
Function | Purpose | Example |
|---|---|---|
| Zero fill karna |
|
| One fill karna |
|
| Custom value fill karna |
|
7. np.eye()
Concept: np.eye() ka use Identity-like matrix create karne ke liye hota hai.
Ismein main diagonal par 1 hota hai aur baaki sabhi positions par 0 hota hai.
Identity matrix kya hoti hai?
Example:
3 × 3 Identity Matrix
1
0
0
0
1
0
0
0
1
Example 1: Basic identity matrix
arr = np.eye(3) # 3x3 matrix create hogi, diagonal par 1 aur baaki jagah 0
print(arr) # Matrix print karengeOutput:
[[1. 0. 0.]
[0. 1. 0.]
[0. 0. 1.]]Example 2: Rectangular matrix
arr = np.eye(3, 4) # 3 rows aur 4 columns ki matrix create hogi
print(arr) # Matrix print karengeOutput:
[[1. 0. 0. 0.]
[0. 1. 0. 0.]
[0. 0. 1. 0.]]Example 3: Diagonal shift karna
np.eye() mein k parameter se diagonal ki position change kar sakte hain.
arr = np.eye(3, k=1) # k=1 se main diagonal ke ek position upar 1 place honge
print(arr) # Matrix print karengeOutput:
[[0. 1. 0.]
[0. 0. 1.]
[0. 0. 0.]]Real-life use: Linear algebra, matrix multiplication, machine learning aur mathematical computations mein identity matrices ka use hota hai.
8. np.identity()
Concept: np.identity() bhi identity matrix create karta hai.
Ye np.eye() jaisa hi hai, lekin ek important difference hai.
np.eye()mein rows aur columns alag-alag specify kar sakte hain.np.identity()mein sirf square matrix ka size specify karte hain.
Example:
arr = np.identity(3) # 3x3 square identity matrix create hogi
print(arr) # Matrix print karengeOutput:
[[1. 0. 0.]
[0. 1. 0.]
[0. 0. 1.]]Integer identity matrix
arr = np.identity(3, dtype=int) # dtype=int se matrix integer values mein create hogi
print(arr) # Matrix print karengeOutput:
[[1 0 0]
[0 1 0]
[0 0 1]]np.eye() vs np.identity()
np.eye() | np.identity() |
|---|---|
Rectangular matrix bana sakta hai | Sirf square matrix banata hai |
Rows aur columns dono specify kar sakte hain | Sirf size specify karte hain |
|
|
More flexible | Simple identity matrix creation |
9. Creating 2D Arrays
Ab samajhte hain ki NumPy mein 2D arrays kaise create karte hain.
Concept: 2D array ek aisa array hota hai jisme rows aur columns hote hain.
Simple language mein, 2D array ko hum table ki tarah samajh sakte hain.
Example 1: Student marks
import numpy as np # NumPy import karenge
marks = np.array([[80, 85, 90], # First student ke 3 subjects ke marks
[70, 75, 88], # Second student ke marks
[90, 92, 95]]) # Third student ke marks
print(marks) # Complete 2D array print karenge
print(marks.ndim) # Array ke dimensions check karenge
print(marks.shape) # Rows aur columns ki information milegiOutput:
[[80 85 90]
[70 75 88]
[90 92 95]]
2
(3, 3)Yahan:
ndim = 2means array 2-dimensional hai.shape = (3, 3)means 3 rows aur 3 columns hain.
Example 2: 2D array indexing
marks = np.array([[80, 85, 90], [70, 75, 88]]) # 2 rows aur 3 columns ka array
print(marks[0, 0]) # First row aur first column ki value
print(marks[0, 1]) # First row aur second column ki value
print(marks[1, 2]) # Second row aur third column ki valueOutput:
80
85
88Remember:
Indexing 0 se start hoti hai.
marks[row, column]se element access karte hain.
10. Creating 3D Arrays
Concept: 3D array mein rows aur columns ke saath ek additional dimension hota hai.
Isko hum multiple 2D matrices ke collection ki tarah samajh sakte hain.
Example: Tumhare paas 2 classrooms hain, har classroom mein 2 students hain aur har student ke 3 subject marks hain.
Example:
import numpy as np # NumPy import karenge
arr = np.array([ # 3D array create karna start karenge
[[80, 85, 90], # Classroom 1: Student 1 ke marks
[70, 75, 88]], # Classroom 1: Student 2 ke marks
[[60, 65, 70], # Classroom 2: Student 1 ke marks
[90, 92, 95]] # Classroom 2: Student 2 ke marks
])
print(arr) # Complete 3D array print karenge
print(arr.ndim) # Dimensions check karenge
print(arr.shape) # Har dimension ka size check karengeOutput:
[[[80 85 90]
[70 75 88]]
[[60 65 70]
[90 92 95]]]
3
(2, 2, 3)Shape (2, 2, 3) ka meaning:
Dimension | Meaning |
|---|---|
First 2 | Total 2 classrooms |
Second 2 | Har classroom mein 2 students |
Third 3 | Har student ke 3 subject marks |
3D array indexing
print(arr[0, 0, 0]) # First classroom, first student, first subject = 80
print(arr[0, 1, 2]) # First classroom, second student, third subject = 88
print(arr[1, 0, 1]) # Second classroom, first student, second subject = 65Output:
80
88
6511. Creating Multidimensional Arrays
Concept: 3D se zyada dimensions wale arrays ko higher-dimensional ya multidimensional arrays kaha jata hai.
NumPy mein hum 4D, 5D aur even higher-dimensional arrays bhi create kar sakte hain.
Example: 4D array
arr = np.ones((2, 3, 4, 5)) # 4D array create hoga jisme dimensions ka size 2, 3, 4, 5 hai
print(arr.ndim) # Total dimensions check karenge
print(arr.shape) # Har dimension ka size check karenge
print(arr.size) # Total elements count karengeOutput:
4
(2, 3, 4, 5)
120Shape (2, 3, 4, 5) ka meaning:
First dimension = 2
Second dimension = 3
Third dimension = 4
Fourth dimension = 5
Total elements:
Different dimensions create karna
arr_1d = np.zeros(5) # 1D array create hoga
arr_2d = np.zeros((2, 3)) # 2D array create hoga
arr_3d = np.zeros((2, 3, 4)) # 3D array create hoga
arr_4d = np.zeros((2, 3, 4, 5)) # 4D array create hoga
print(arr_1d.ndim) # Output: 1
print(arr_2d.ndim) # Output: 2
print(arr_3d.ndim) # Output: 3
print(arr_4d.ndim) # Output: 4Important NumPy attributes
Attribute | Meaning | Example |
|---|---|---|
| Total dimensions |
|
| Har dimension ka size |
|
| Total elements |
|
| Elements ka data type |
|
Ye four attributes NumPy arrays ke saath kaafi frequently use hote hain.
12. Understanding dtype
Concept: dtype ka full form hai Data Type.
Ye batata hai ki NumPy array ke andar kis type ka data stored hai.
For example:
Integer numbers
Decimal numbers
Boolean values
Complex numbers
Example 1: Integer array
arr = np.array([10, 20, 30, 40]) # Integer values ka array create karenge
print(arr.dtype) # Array ka data type check karengePossible output:
int64Example 2: Float array
arr = np.array([10.5, 20.5, 30.5]) # Decimal values ka array create karenge
print(arr.dtype) # Data type check karengeOutput:
float64Example 3: String array
arr = np.array(["Apple", "Mango", "Banana"]) # String values ka array create karenge
print(arr.dtype) # String data type check karengeOutput: String ke liye NumPy Unicode dtype show karega, jaise <U6.
Common NumPy data types
Data type | Meaning | Example |
|---|---|---|
| 32-bit integer | 100 |
| 64-bit integer | 100 |
| 32-bit decimal | 10.5 |
| 64-bit decimal | 10.5 |
| True / False | True |
| Complex number | 2+3j |
Important: Data type memory usage aur calculations ki precision ko affect karta hai.
13. Creating Arrays Using dtype
Hum array create karte waqt hi uska data type specify kar sakte hain.
Example 1: Integer dtype
arr = np.array([1.5, 2.5, 3.5], dtype=int) # Values ko integer dtype mein store karenge
print(arr) # Converted array print karenge
print(arr.dtype) # Data type check karengeOutput:
[1 2 3]
int64Notice karo decimal part remove ho gaya.
Example 2: Float dtype
arr = np.array([1, 2, 3], dtype=float) # Integer values ko float type mein create karenge
print(arr) # Array print karenge
print(arr.dtype) # Data type check karengeOutput:
[1. 2. 3.]
float64Example 3: Boolean dtype
arr = np.array([1, 0, 1, 0], dtype=bool) # Values ko Boolean type mein convert karenge
print(arr) # Array print karengeOutput:
[ True False True False]14. Type Conversion Using astype()
Concept: Kabhi-kabhi array already create ho chuka hota hai, lekin humein uska data type change karna hota hai.
Iske liye NumPy mein astype() method use hota hai.
Syntax:
new_array = old_array.astype(new_dtype)Example 1: Integer to Float
arr = np.array([10, 20, 30]) # Integer array create karenge
new_arr = arr.astype(float) # Integer array ko float array mein convert karenge
print(arr) # Original array print karenge
print(new_arr) # Converted array print karenge
print(new_arr.dtype) # Converted array ka data type check karengeOutput:
[10 20 30]
[10. 20. 30.]
float64Yahan original array change nahi hua. astype() ne ek naya converted array return kiya.
Example 2: Float to Integer
arr = np.array([10.8, 20.5, 30.9]) # Float array create karenge
new_arr = arr.astype(int) # Float values ko integer mein convert karenge
print(new_arr) # Converted array print karengeOutput:
[10 20 30]Important: astype(int) decimal values ko round off nahi karta. Ye decimal part truncate karta hai.
Example 3: Integer to String
arr = np.array([10, 20, 30]) # Integer array create karenge
new_arr = arr.astype(str) # Integer values ko string mein convert karenge
print(new_arr) # Converted array print karenge
print(new_arr.dtype) # Data type check karengeOutput:
['10' '20' '30']
<U21dtype vs astype()
dtype | astype() |
|---|---|
Array create karte waqt use hota hai | Existing array ka type convert karta hai |
Initial data type decide karta hai | New data type mein converted array deta hai |
|
|
15. All Functions – Quick Revision
Ab tak humne jitne functions seekhe, unhe ek jagah revise karte hain.
Function | Purpose | Example |
|---|---|---|
| Sequence generate karna |
|
| Equally spaced values |
|
| Zero array |
|
| One array |
|
| Custom value array |
|
| Identity-like matrix |
|
| Square identity matrix |
|
| Data type conversion |
|
16. Practical Mini Project – Student Marks Dataset
Ab ek practical example banate hain jisme humne seekhe hue concepts ko combine karenge.
Scenario: Ek school mein 3 students hain aur unke 3 subjects ke marks store karne hain.
import numpy as np # NumPy import karenge
marks = np.array([ # Students ke marks ka 2D array create karenge
[80, 85, 90], # Student 1 ke marks
[70, 75, 88], # Student 2 ke marks
[90, 92, 95] # Student 3 ke marks
], dtype=int) # Marks ko integer data type mein store karenge
print(marks) # Complete marks dataset print karenge
print(marks.shape) # Rows aur columns check karenge
print(marks.ndim) # Dimensions check karenge
print(marks.dtype) # Data type check karenge
average = np.mean(marks, axis=1) # Har student ke subjects ka average calculate karenge
print(average) # Har student ka average print karenge
new_marks = np.full((3, 3), 5) # Har student ke har subject ke liye 5 bonus marks create karenge
final_marks = marks + new_marks # Original marks mein bonus marks add karenge
print(final_marks) # Final marks print karenge
float_marks = final_marks.astype(float) # Final marks ko float type mein convert karenge
print(float_marks.dtype) # Converted array ka dtype check karengeOutput:
[[80 85 90]
[70 75 88]
[90 92 95]]
(3, 3)
2
int64
[85. 77.66666667 92.33333333]
[[ 85 90 95]
[ 75 80 93]
[ 95 97 100]]
float64Is ek example mein humne array creation, dimensions, dtype, np.full(), mathematical operations aur type conversion sab use kiya.
17. Practice Questions
Ab tumhari understanding check karte hain. Pehle khud solve karna, phir answers check karna.
NumPy Knowledge Test
1 se 20 tak even numbers ka NumPy array kaise create karoge?
0 se 100 ke beech exactly 6 equally spaced values ke liye kaunsa function use hoga?
3 rows aur 4 columns ka zero array kaise create karoge?
Existing array ko float mein convert karne ke liye kya use karoge?
np.identity(4) ka output shape kya hoga?
18. Important Points to Remember
np.arange()→ Step size define karte hain.np.linspace()→ Total numbers define karte hain.np.zeros()→ All elements zero.np.ones()→ All elements one.np.full()→ All elements custom value.np.eye()→ Diagonal matrix.np.identity()→ Square identity matrix.ndim→ Dimensions count.shape→ Dimensions ka size.dtype→ Data type check ya define karna.astype()→ Existing array ka data type convert karna.
My recommendation: NumPy mein advanced array creation achhe se master karne ke baad next topic Array Indexing, Slicing and Iteration seekhna. Ye NumPy aur Pandas dono ke liye bahut important foundation hai.
Comments
Post a Comment