import numpy as np

# Check version
print(np.__version__)

import numpy as np

arr = np.array([1,2,3,4,5])
# np.array((1,2,3,4,5)), Can pass a list, it will be converted into an ndarray
print(arr)	# output [1 2 3 4 5]
print(type(arr))	# output <class ‘numpy.ndarray’>
print(arr.ndim)	# output 1

# Array can have any number of dimentions
arr2 = np.array(arr,ndmin=5)		# output [[[[[1 2 3 4 5]]]]]

# Access Element, Is same as accessing an array element
print(arr[0])	# output 1

# Access elements from 2D arrays
arr3 = np.array([[1,2,3],[4,5,6]])
print(arr[0,1])		# output 2

# Negative indexing
print(arr[1,-1])		# output 6

# Slicing
print(arr[1:3])		# output [2 3]
print(arr[3:])		# output [4 5]
print(arr[-3:-1])	# output [3 4]
print(arr[1:4:2])	# output [2 4]
print(arr[::2])		# output [1 3 5]

# Slicing 2D
print(arr3[1, 1:3])	# output [5 6]
print(arr3[0:2, 2])	# output [3 6]
print(arr3[:3, 1:3])	# output [[2 3] [5 6]]

import numpy as np

arr = np.array([‘apple’, ‘banana’, ‘cherry’])
# Check datatype
print(arr.dtype)	# output <U6

# Defined data type
arr2 = np.array([1,2,3,4], dtype=‘S’)
print(arr2)	# output [b‘1’ b‘2’ b‘3’ b‘4’]
print(arr2.dtype)	# output |S1

# i, u, f, S, U can define size, if can not be converted, Error is raised
arr3 = np.array([1.2, 2.3, 3.4, 4.5], dtype=‘f4’)
print(arr3)		# output [1.2 2.3 3.4 4.5]
print(arr.dtype)	# output float32

# Converting data type
arr4 = arr3.astype(‘i’)	# or can use the data type directly like int for integer
print(arr4)		# output [1 2 3 4]
print(arr4.dtype)	# output int32

import numpy as np

arr = np.array([1,2,3,4,5])
x = arr.copy()
y = arr.view()
arr[0] = 42
print(arr)	# output [42 2 3 4 5]
print(x)	# output [42 2 3 4 5]
print(x.base)	# output [42 2 3 4 5]
print(y)	# output [1 2 3 4 5]
print(y.base)	# output None

import numpy as np

arr = np.array([[1,2,3,4],[5,6,7,8]])

# Returns a tuple with each index having the number of corresponding elements
print(arr.shape)	# output (2,4)

# Can add or remove dimensions or change number of elements in each dimension
newarr = arr.reshape(8)
print(newarr)		# output [1 2 3 4 5 6 7 8]
# It is a view
print(newarr.base)	# output [[1 2 3 4] [5 6 7 8]]
# Unknown Dimension, numpy will calculate this number, can’t pass -1 to more than one dimension
print(arr.reshape(2,2,-1))	# output [[[1 2] [3 4] [5 6] [7 8]]]
# Flattening, converting a multidimentional array into a 1D
print(arr.reshape(-1))		# output [1 2 3 4 5 6 7 8]

# Iterate on a n-D array it will go through n-1th dimension one by one
# To return the actual values, the scalars, we have to iterate the arrays in each dimension
import numpy as np

arr = np.array([[[1,2,3],[4,5,6]], [[7,8,9],[10,11,12]]])
for a in arr:
    for b in a:
        for c in b:
            print(z)
# helping function that can be used iterations
# Both are same result
for x in np.nditer(arr):
    print(x)

# op_dtypes arg can pass it the expected datatype to change the datatype of elements while iterating
for x in np.nditer(arr, flags=[‘buffered’], op_dtypes=[‘S’]):
    print(x, end= ‘ ’)	# output b‘1’ b‘2’ b‘3’
# Numpy doesn’t change the data type of the element in-place,
# So needs some other space(buffer) to change dtype

# Different step
for x in np.nditer(arr[:,::2]):
    print(x, end= ‘ ’)	# output 1 2 3 7 8 9

# Enumerated iteration
for idx, x in np.ndenumerate(arr):
    print(idx, x, end= ‘ ’)	# output (0,0,0) 1 (0,0,1) 2 …(1,0,0) 7…(1,1,2) 12

import numpy as np

arr1 = np.array([[1,2],[3,4]])
arr2 = np.array([[5,6],[7,8]])
arr3 = np.concatenate((arr1,arr2), axis = 1)	# output [[1 2 5 6] [3 4 7 8]]
arr4 = np.concatenate((arr1,arr2))	# output [[1 2] [3 4] [5 6] [7 8]]

# stack : only difference is that stacking is done along a new axis
arr5 = np.stack((arr1,arr2), axis=1)	# output [[[1 2] [5 6]] [[3 4] [7 8]]]

# To stack along rows
arr6 = np.hstack((arr1,arr2))	# output [[1 2 5 6] [3 4 7 8]]

# To stack along columns
arr7 = np.vstack((arr1,arr2))	# output [[1 2] [3 4] [5 6] [7 8]]

# To stack along height
arr8 = np.dstakc((arr1,arr2))	# output [[[1 5] [2 6]] [[3 7] [4 8]]]

import numpy as np

arr = np.array([[1,2], [3,4], [5,6], [7,8]])
arr1 = np.array_split(arr,2)	# output [array([[1, 2], [3,4]]), array([[5,6],[7,8]])]
# If the array has less elements than required, it will adjust from the end accordingly
arr2 = np.array_split(arr,3)	# output [array([[1,2], [3,4]]), array([[5,6]]), array([[7,8]])]
# Can specify which axis you want to do the split around
arr3 = np.array_split(arr,2,axis=1)	# output [array([[1], [3], [5], [7]]), array([[2], [4], [6], [8]])]
# An alternate solution is hsplit()
# Similar alternates to vstack() and dstack() are available as vsplit() and dsplit()

import numpy as np

arr = np.array([1,2,3,4,5,4,4])
a = np.where(arr==4)
print(a)	# output (array([3,5,6]),)

# Performs a binary search in the array, and returns the index where the specified value would be inserted to maintain the search order
arr1 = np.array([6,7,8,9])
b = np.searchsorted(arr1, 7)
print(b)	# output 1

# To return the right most index instead
c = np.searchsorted(arr1, 7, side= ‘right’)
print(c)		# output 2

# To search for more than one value, use an array with the specified values
d = np.searchsorted(arr1, [1,2,10])
print(d)	# output [0 0 4]

import numpy as np

arr = np.array([41,42,43,44])
a = [True, False, True, False]
arr1 = arr[a]
print(arr1)	# output  [41 43]

# Create filter directly
filter_arr = arr > 42
arr2 = arr[filter_arr]
print(filter_arr)	# output [False Flase True True]
print(arr2)	# output [43 44]