Indexing

One dimensional array

Omitting the start index starts the slice from the index 0.

Meaning, A[: stop] is equivalent to A[0:stop]

Omitting the stop index extends the slice to the end of the array.

Meaning A[start:] is equivalent ot A[start:len(A)]

Two dimensional array

In 2D array, the elements at each index are no longer scalars, but rather one-dimensional array

np.array([[10,20,30], [40,50,60], [70,80,90]])

Multidimensional arrays can have one index per axis. These indices are given in a tuple separated by commas

Slice – selects a range of elements along an axis

Multiple slices can be passed as like multiple index

b[:2] – Slicing along row ( 2 exclusive)

fromfunction( )

construct an array by executing a function over each coordinate and the resulting array,

def f(x,y):
    return 10*x+y
b1 = np.fromfunction(f,(5,4),dtype=int)
b1

array([[ 0,  1,  2,  3],
       [10, 11, 12, 13],
       [20, 21, 22, 23],
       [30, 31, 32, 33],
       [40, 41, 42, 43]])

#getting particular element
print("b1[2,3]=",b1[2,3])
#Slicing and indexig
print("b1[0:3,2]=",b1[0:3,2])
print("b1[:,1]=",b1[:,1])#first column
print("b1[1:3,:]=",b1[1:3,:])

b1[2,3]= 23
b1[0:3,2]= [ 2 12 22]
b1[:,1]= [ 1 11 21 31 41]
b1[1:3,:]= [[10 11 12 13]
 [20 21 22 23]]

Multi Dimensional Array

In Multidimensional arrays, If later indices are omitted, the returned object will be a lower dimensional ndarray consisting of all the data along the higher dimensions

Boolean Indexing

Passing Boolean array as index

narr1 = np.array([10,20,30,40])
narr2 = np.array([False,True,False, True])
print("Selected elements:",narr1[narr2])

Selected elements: [20 40]

Selecting even numbers based on Boolean array

ba1=np.empty(10)
for i in range(10):
    ba1[i] = i
print(ba1)
ba2 = (ba1 % 2 == 0)
print(ba2)
print("even numbers:",ba1[ba2])

[0. 1. 2. 3. 4. 5. 6. 7. 8. 9.]
[ True False  True False  True False  True False  True False]
even numbers: [0. 2. 4. 6. 8.]

data

ba3 = np.array(['ram','abi','ravi','ram','ravi'])
data = np.random.randn(5,4)
print(data)
ba4 = (ba3 =='ram')
print("ba4",ba4)
#search for specific name
print(data [ ba3 == 'ram']) # boolean array can be passed when indexing the array:

[[-1.15594765  1.30522028 -2.49459926  0.39999895]
 [ 1.47418789 -0.1766457   0.39876725 -0.56476218]
 [ 1.55699037  1.30027161 -1.21973007  0.36258873]
 [ 0.03846156  0.44934939 -0.25101183  2.13427514]
 [-0.38013597 -0.36425022  0.42257762 -0.01005772]]
ba4 [ True False False  True False]
[[-1.15594765  1.30522028 -2.49459926  0.39999895]
 [ 0.03846156  0.44934939 -0.25101183  2.13427514]]

#boolean array must be of the same length as the axis it’s indexing. even mix and match boolean arrays with slices or integers
print("data[ba3 == 'abi',2]",data[ba3 == 'abi',2])
print("data[ba3 == 'abi',2:]",data[ba3 == 'abi',2:])

data[ba3 == 'abi',2] [0.39876725]
data[ba3 == 'abi',2:] [[ 0.39876725 -0.56476218]]

Select everything but not ram

#Select everything but not ram
print("(ba3 != 'ram') = ",ba3 != 'ram')

(ba3 != 'ram') =  [False  True  True False  True]

Selecting two of the three names to combine multiple boolean conditions, use boolean arithmetic operators like & (and) and | (or)

#Selecting two of the three names to combine multiple boolean conditions, use boolean arithmetic operators like & (and) and | (or):
mask = ( ba3 == 'ram') | (ba3 == 'abi')
print("mask=",mask)

mask= [ True  True False  True False]

Setting values with boolean array

set all of the negative values in data to 0 
data[data<0] =0
print("Now data=",data)

Now data= [[0.         1.30522028 0.         0.39999895]
 [1.47418789 0.         0.39876725 0.        ]
 [1.55699037 1.30027161 0.         0.36258873]
 [0.03846156 0.44934939 0.         2.13427514]
 [0.         0.         0.42257762 0.        ]]

Setting whole rows or columns using 1D boolean array

data[ba3 != 'abi'] = 1
print(data)

[[1.         1.         1.         1.        ]
 [1.47418789 0.         0.39876725 0.        ]
 [1.         1.         1.         1.        ]
 [1.         1.         1.         1.        ]
 [1.         1.         1.         1.        ]]

Fancy Indexing

Base Array:

for i in range(8):
    arr[i] = i
print(arr)

[[0. 0. 0. 0.]
 [1. 1. 1. 1.]
 [2. 2. 2. 2.]
 [3. 3. 3. 3.]
 [4. 4. 4. 4.]
 [5. 5. 5. 5.]
 [6. 6. 6. 6.]
 [7. 7. 7. 7.]]

Select a subset of rows in particular pass a list or ndarray of integers in desired order (arr)

print(arr[[4,3,1,6]])

Output

[[4. 4. 4. 4.]
 [3. 3. 3. 3.]
 [1. 1. 1. 1.]
 [6. 6. 6. 6.]]

Use negative indexing to select from the end

print(arr[[-2,-3,-1,-5]])

Output

[[6. 6. 6. 6.]
 [5. 5. 5. 5.]
 [7. 7. 7. 7.]
 [3. 3. 3. 3.]]

Passing multiple index arrays

#Passing multiple index arrays
arr1 = np.arange(32).reshape((8,4))
print(arr1)

[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]
 [12 13 14 15]
 [16 17 18 19]
 [20 21 22 23]
 [24 25 26 27]
 [28 29 30 31]]

Selects a 1D array of elements corresponding to each tuple of indices

#Selects a 1D array of elements corresponding to each tuple of indices
print(arr1[ [1,4,2,6],[2,3,1,0] ])
# the elements at (1,2) , (4,3), (2,1) and (6,0) are selected

Output

[ 6 19  9 24]

To select the rectangular region formed by selecting a subset of matrix’s row and column

print(arr1[ [1,4,2,6] ] [:,[2,3,1,0]] )
the elements at the position (1,2); (1,3);(1,1);(1,0) are selected for first row and so on

Output

[[ 6  7  5  4]
 [18 19 17 16]
 [10 11  9  8]
 [26 27 25 24]]

Function converts two 1D integer arrays to an indexer that selects the square region

np_ix_ function converts two 1D integer arrays to an indexer that selects the square region
print(arr1[np.ix_( [1,4,2,6], [2,3,1,0] )])
the elements at the position (1,2); (1,3);(1,1);(1,0) are selected for first row and so on

Output

[[ 6  7  5  4]
 [18 19 17 16]
 [10 11  9  8]
 [26 27 25 24]]

Views: 0