![\"NumPy](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2020\/07\/Arrays-in-NumPy-df.jpg\")
<\/a><\/p>\nNumpy Array Basics<\/h2>\n
Arrays in NumPy are synonymous with lists in Python with a homogenous nature. The homogeneity helps to perform smoother mathematical operations.<\/p>\n
These arrays are mutable. NumPy is useful to perform basic operations like finding the dimensions, the bite-size, and also the data types of elements of the array.<\/p>\n
NumPy Array Creation<\/h2>\n![\"Array](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2020\/07\/Array-Creation-dF.jpg\")
<\/a><\/h3>\n1. Using the NumPy functions<\/h3>\n
NumPy has a variety of built-in functions to create an array.<\/p>\n
a. Creating one-dimensional array in NumPy<\/h4>\n
For 1-D arrays the most common function is np.arange(..), passing any value create an array from 0 to that number.<\/p>\n
import numpy as np\r\n array=np.arange(20)\r\n array\r\n<\/pre>\nOutput<\/strong><\/p>\narray([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,12, 13, 14, 15, 16, 17, 18, 19])<\/div>\nWe can check the dimensions by using array.shape.<\/strong><\/p>\nOutput<\/strong><\/p>\n(20,)<\/div>\nTo access the element of the array we can specify its non-negative index.
\narray[3]<\/strong><\/p>\nOutput<\/strong><\/p>\n3<\/div>\nb. Creating two-dimensional arrays in NumPy<\/h4>\n
We can use reshape()function along with the arange() function to create a 2D array. The reshape()function specifies the rows and columns.<\/p>\n
array=np.arange(20).reshape(4,5)<\/pre>\nOutput<\/strong><\/p>\narray([[ 0, 1, 2, 3, 4],
\n[ 5, 6, 7, 8, 9],
\n[10, 11, 12, 13, 14]
\n[15, 16, 17, 18, 19]])<\/div>\nIn 2-D arrays to access the elements, we will require to specify two values for row and column respectively.<\/p>\n
Similarly, we can create 3-D and more by increasing the number of parameters in the reshape()function.<\/p>\n
c. Using other NumPy functions<\/h4>\n
We can use other functions like and to quickly create filled arrays.<\/p>\n
np.zeros((2,4))\r\n np.ones((3,6))<\/pre>\nOutput<\/strong><\/p>\narray([[0., 0., 0., 0.],
\n[0., 0., 0., 0.]])array([[1., 1., 1., 1., 1., 1.],
\n[1., 1., 1., 1., 1., 1.],
\n[1., 1., 1., 1., 1., 1.]])<\/div>\nThe .empty() function creates an array with random variables and the full() function creates an n*n array with the given value.<\/p>\n
np.empty((2,3))\r\n np.full((2,2), 3)<\/pre>\nOutput<\/strong><\/p>\narray([[2.69893675e-316, 0.00000000e+000, 0.00000000e+000],
\n[0.00000000e+000, 0.00000000e+000, 0.00000000e+000]])
\narray([[3, 3],
\n[3, 3]])<\/div>\nThe .eye( , )function creates an array with diagonals as 1 and other
\nvalues as 0. The .linspace(, ,)
\nfunction outputs an equally spaced array.<\/p>\n
np.eye(3,3)\r\n np.linspace(0, 10, num=4)<\/pre>\nOutput<\/strong><\/p>\narray([[1., 0., 0.],
\n[0., 1., 0.],
\n[0., 0., 1.]])
\narray([ 0. , 3.33333333, 6.66666667, 10. ])<\/div>\n\n\n\nSr No.<\/strong><\/td>\nFunction<\/strong><\/td>\nDescription<\/strong><\/td>\n<\/tr>\n\n1<\/span><\/td>\n empty_like()<\/span><\/td>\n Return a new array with the same shape and type\u00a0<\/span><\/td>\n<\/tr>\n\n2<\/span><\/td>\n ones_like()<\/span><\/td>\n Return an array of ones with the same shape and type.<\/span><\/td>\n<\/tr>\n\n3<\/span><\/td>\n zeros_like()<\/span><\/td>\n Return an array of zeros with the same shape and type\u00a0<\/span><\/td>\n<\/tr>\n\n4<\/span><\/td>\n full_like()<\/span><\/td>\n Return a full array with the same shape and type\u00a0<\/span><\/td>\n<\/tr>\n\n5<\/span><\/td>\n asarray()<\/span><\/td>\n Convert the input to an array.<\/span><\/td>\n<\/tr>\n\n6<\/span><\/td>\n geomspace()<\/span><\/td>\n Return evenly spaced numbers on a log scale.<\/span><\/td>\n<\/tr>\n\n7<\/span><\/td>\n copy()<\/span><\/td>\n Returns a copy of the given object<\/span><\/td>\n<\/tr>\n\n8<\/span><\/td>\n diag()<\/span><\/td>\n a diagonal array<\/span><\/td>\n<\/tr>\n\n9<\/span><\/td>\n frombuffer()<\/span><\/td>\n buffer as a 1-D array<\/span><\/td>\n<\/tr>\n\n10<\/span><\/td>\n fromfile()<\/span><\/td>\n Construct an array from text or binary file<\/span><\/td>\n<\/tr>\n\n11<\/span><\/td>\n bmat()<\/span><\/td>\n Build a matrix object from a string, nested sequence, or array<\/span><\/td>\n<\/tr>\n\n12<\/span><\/td>\n mat()<\/span><\/td>\n Interpret the input as a matrix<\/span><\/td>\n<\/tr>\n\n13<\/span><\/td>\n vander()<\/span><\/td>\n Generate a Vandermonde matrix<\/span><\/td>\n<\/tr>\n\n14<\/span><\/td>\n triu()<\/span><\/td>\n Upper triangle of array<\/span><\/td>\n<\/tr>\n\n15<\/span><\/td>\n tril()<\/span><\/td>\n Lower triangle of array<\/span><\/td>\n<\/tr>\n\n16<\/span><\/td>\n tri()<\/span><\/td>\n An array with ones at & below the given diagonal and zeros elsewhere<\/span><\/td>\n<\/tr>\n\n17<\/span><\/td>\n diagflat()<\/span><\/td>\n two-dimensional array with the flattened input as a diagonal<\/span><\/td>\n<\/tr>\n\n18<\/span><\/td>\n fromfunction()<\/span><\/td>\n executing a function over each coordinate<\/span><\/td>\n<\/tr>\n\n19<\/span><\/td>\n logspace()<\/span><\/td>\n Return numbers spaced evenly on a log scale<\/span><\/td>\n<\/tr>\n\n20<\/span><\/td>\n meshgrid()<\/span><\/td>\n Return coordinate matrices from coordinate vectors<\/span><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n2. Conversion from Python structure like lists<\/h3>\n
We can use the Python lists to create arrays by passing a list to the array function. We can also directly create an array of elements by passing a list.<\/p>\n
array=np.array([4,5,6])\r\n array\r\n \r\n list=[4,5,6]\r\n list\r\n<\/pre>\nOutput<\/strong><\/p>\n\narray([4, 5, 6])[4, 5, 6]<\/p>\n<\/div>\n
To create 2-D or more we pass a sequence of lists.<\/p>\n
3. Using other library functions<\/h3>\n
We can use the function to create an array with random values between 0 and 1. This is a useful case in scenarios with random nature.<\/p>\n
np.random.random((2,3))<\/pre>\nOutput<\/strong><\/p>\narray([[0.42593659, 0.91495384, 0.05727104],
\n[0.3754818 , 0.63166016, 0.5901392 ]])<\/div>\nNumPy Array Indexing<\/h2>\n
Indexing of the array has to be proper in order to access and manipulate its values. Indexing can be done through:<\/p>\n
\n- Slicing<\/strong> \u2013 we perform slicing on NumPy arrays with the declaration of a slice for all the dimensions.<\/li>\n
- Integer array Indexing<\/strong>– users can pass lists for one to one mapping of corresponding elements for each dimension.<\/li>\n
- Boolean Array Indexing<\/strong>– we can pick elements after satisfying a particular Boolean condition.<\/li>\n<\/ul>\n
NumPy Basic Array Operations<\/h2>\n
There is a vast range of built-in operations that we can perform on these arrays.<\/p>\n
1. ndim<\/strong> – It returns the dimensions of the array.
\n2. itemsize<\/strong> \u2013 It calculates the byte size of each element.
\n3. dtype<\/strong> – It can determine the data type of the element.
\n4. reshape<\/strong> – It provides a new view.
\n5. slicing<\/strong> – It extracts a particular set of elements.
\n6. linspace<\/strong> – Returns evenly spaced elements.
\n7.<\/strong> max\/min , sum, sqrt
\n8. ravel<\/strong> – It converts the array into a single line.<\/p>\nThere are also a few Special Operations like sine, cosine, tan, log, etc.<\/p>\n
Checking Array Dimensions in NumPy<\/h2>\n
We can determine the NumPy array dimensions using the ndim attribute. The argument return an integer that indicates the array dimension.<\/p>\n
import numpy as np\r\n\r\na = np.array(10)\r\nb = np.array([1,1,1,1])\r\nc = np.array([[1, 1, 1], [2,2,2]])\r\nd = np.array([[[1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4]]])\r\n\r\nprint(a.ndim)\r\nprint(b.ndim)\r\nprint(c.ndim)\r\nprint(d.ndim)\r\n\r\n<\/pre>\nOutput<\/strong><\/p>\n
<\/a><\/h3>\n1. Using the NumPy functions<\/h3>\n
NumPy has a variety of built-in functions to create an array.<\/p>\n
a. Creating one-dimensional array in NumPy<\/h4>\n
For 1-D arrays the most common function is np.arange(..), passing any value create an array from 0 to that number.<\/p>\n
import numpy as np\r\n array=np.arange(20)\r\n array\r\n<\/pre>\nOutput<\/strong><\/p>\n
array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,12, 13, 14, 15, 16, 17, 18, 19])<\/div>\nWe can check the dimensions by using array.shape.<\/strong><\/p>\n
Output<\/strong><\/p>\n
(20,)<\/div>\nTo access the element of the array we can specify its non-negative index.
\narray[3]<\/strong><\/p>\nOutput<\/strong><\/p>\n
3<\/div>\nb. Creating two-dimensional arrays in NumPy<\/h4>\n
We can use reshape()function along with the arange() function to create a 2D array. The reshape()function specifies the rows and columns.<\/p>\n
array=np.arange(20).reshape(4,5)<\/pre>\nOutput<\/strong><\/p>\n
array([[ 0, 1, 2, 3, 4],
\n[ 5, 6, 7, 8, 9],
\n[10, 11, 12, 13, 14]
\n[15, 16, 17, 18, 19]])<\/div>\nIn 2-D arrays to access the elements, we will require to specify two values for row and column respectively.<\/p>\n
Similarly, we can create 3-D and more by increasing the number of parameters in the reshape()function.<\/p>\n
c. Using other NumPy functions<\/h4>\n
We can use other functions like and to quickly create filled arrays.<\/p>\n
np.zeros((2,4))\r\n np.ones((3,6))<\/pre>\nOutput<\/strong><\/p>\n
array([[0., 0., 0., 0.],
\n[0., 0., 0., 0.]])array([[1., 1., 1., 1., 1., 1.],
\n[1., 1., 1., 1., 1., 1.],
\n[1., 1., 1., 1., 1., 1.]])<\/div>\nThe .empty() function creates an array with random variables and the full() function creates an n*n array with the given value.<\/p>\n
np.empty((2,3))\r\n np.full((2,2), 3)<\/pre>\nOutput<\/strong><\/p>\n
array([[2.69893675e-316, 0.00000000e+000, 0.00000000e+000],
\n[0.00000000e+000, 0.00000000e+000, 0.00000000e+000]])
\narray([[3, 3],
\n[3, 3]])<\/div>\nThe .eye( , )function creates an array with diagonals as 1 and other
\nvalues as 0. The .linspace(, ,)
\nfunction outputs an equally spaced array.<\/p>\nnp.eye(3,3)\r\n np.linspace(0, 10, num=4)<\/pre>\nOutput<\/strong><\/p>\n
array([[1., 0., 0.],
\n[0., 1., 0.],
\n[0., 0., 1.]])
\narray([ 0. , 3.33333333, 6.66666667, 10. ])<\/div>\n\n\n
\n Sr No.<\/strong><\/td>\n Function<\/strong><\/td>\n Description<\/strong><\/td>\n<\/tr>\n \n 1<\/span><\/td>\n empty_like()<\/span><\/td>\n Return a new array with the same shape and type\u00a0<\/span><\/td>\n<\/tr>\n \n 2<\/span><\/td>\n ones_like()<\/span><\/td>\n Return an array of ones with the same shape and type.<\/span><\/td>\n<\/tr>\n \n 3<\/span><\/td>\n zeros_like()<\/span><\/td>\n Return an array of zeros with the same shape and type\u00a0<\/span><\/td>\n<\/tr>\n \n 4<\/span><\/td>\n full_like()<\/span><\/td>\n Return a full array with the same shape and type\u00a0<\/span><\/td>\n<\/tr>\n \n 5<\/span><\/td>\n asarray()<\/span><\/td>\n Convert the input to an array.<\/span><\/td>\n<\/tr>\n \n 6<\/span><\/td>\n geomspace()<\/span><\/td>\n Return evenly spaced numbers on a log scale.<\/span><\/td>\n<\/tr>\n \n 7<\/span><\/td>\n copy()<\/span><\/td>\n Returns a copy of the given object<\/span><\/td>\n<\/tr>\n \n 8<\/span><\/td>\n diag()<\/span><\/td>\n a diagonal array<\/span><\/td>\n<\/tr>\n \n 9<\/span><\/td>\n frombuffer()<\/span><\/td>\n buffer as a 1-D array<\/span><\/td>\n<\/tr>\n \n 10<\/span><\/td>\n fromfile()<\/span><\/td>\n Construct an array from text or binary file<\/span><\/td>\n<\/tr>\n \n 11<\/span><\/td>\n bmat()<\/span><\/td>\n Build a matrix object from a string, nested sequence, or array<\/span><\/td>\n<\/tr>\n \n 12<\/span><\/td>\n mat()<\/span><\/td>\n Interpret the input as a matrix<\/span><\/td>\n<\/tr>\n \n 13<\/span><\/td>\n vander()<\/span><\/td>\n Generate a Vandermonde matrix<\/span><\/td>\n<\/tr>\n \n 14<\/span><\/td>\n triu()<\/span><\/td>\n Upper triangle of array<\/span><\/td>\n<\/tr>\n \n 15<\/span><\/td>\n tril()<\/span><\/td>\n Lower triangle of array<\/span><\/td>\n<\/tr>\n \n 16<\/span><\/td>\n tri()<\/span><\/td>\n An array with ones at & below the given diagonal and zeros elsewhere<\/span><\/td>\n<\/tr>\n \n 17<\/span><\/td>\n diagflat()<\/span><\/td>\n two-dimensional array with the flattened input as a diagonal<\/span><\/td>\n<\/tr>\n \n 18<\/span><\/td>\n fromfunction()<\/span><\/td>\n executing a function over each coordinate<\/span><\/td>\n<\/tr>\n \n 19<\/span><\/td>\n logspace()<\/span><\/td>\n Return numbers spaced evenly on a log scale<\/span><\/td>\n<\/tr>\n \n 20<\/span><\/td>\n meshgrid()<\/span><\/td>\n Return coordinate matrices from coordinate vectors<\/span><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n 2. Conversion from Python structure like lists<\/h3>\n
We can use the Python lists to create arrays by passing a list to the array function. We can also directly create an array of elements by passing a list.<\/p>\n
array=np.array([4,5,6])\r\n array\r\n \r\n list=[4,5,6]\r\n list\r\n<\/pre>\nOutput<\/strong><\/p>\n
\narray([4, 5, 6])[4, 5, 6]<\/p>\n<\/div>\n
To create 2-D or more we pass a sequence of lists.<\/p>\n
3. Using other library functions<\/h3>\n
We can use the function to create an array with random values between 0 and 1. This is a useful case in scenarios with random nature.<\/p>\n
np.random.random((2,3))<\/pre>\nOutput<\/strong><\/p>\n
array([[0.42593659, 0.91495384, 0.05727104],
\n[0.3754818 , 0.63166016, 0.5901392 ]])<\/div>\nNumPy Array Indexing<\/h2>\n
Indexing of the array has to be proper in order to access and manipulate its values. Indexing can be done through:<\/p>\n
\n
- Slicing<\/strong> \u2013 we perform slicing on NumPy arrays with the declaration of a slice for all the dimensions.<\/li>\n
- Integer array Indexing<\/strong>– users can pass lists for one to one mapping of corresponding elements for each dimension.<\/li>\n
- Boolean Array Indexing<\/strong>– we can pick elements after satisfying a particular Boolean condition.<\/li>\n<\/ul>\n
NumPy Basic Array Operations<\/h2>\n
There is a vast range of built-in operations that we can perform on these arrays.<\/p>\n
1. ndim<\/strong> – It returns the dimensions of the array.
\n2. itemsize<\/strong> \u2013 It calculates the byte size of each element.
\n3. dtype<\/strong> – It can determine the data type of the element.
\n4. reshape<\/strong> – It provides a new view.
\n5. slicing<\/strong> – It extracts a particular set of elements.
\n6. linspace<\/strong> – Returns evenly spaced elements.
\n7.<\/strong> max\/min , sum, sqrt
\n8. ravel<\/strong> – It converts the array into a single line.<\/p>\nThere are also a few Special Operations like sine, cosine, tan, log, etc.<\/p>\n
Checking Array Dimensions in NumPy<\/h2>\n
We can determine the NumPy array dimensions using the ndim attribute. The argument return an integer that indicates the array dimension.<\/p>\n
import numpy as np\r\n\r\na = np.array(10)\r\nb = np.array([1,1,1,1])\r\nc = np.array([[1, 1, 1], [2,2,2]])\r\nd = np.array([[[1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4]]])\r\n\r\nprint(a.ndim)\r\nprint(b.ndim)\r\nprint(c.ndim)\r\nprint(d.ndim)\r\n\r\n<\/pre>\nOutput<\/strong><\/p>\n