Numpy learning

numpy A preliminary understanding of

1. Use numpy Generating arrays , obtain ndarray The type of

t1 = np.array([1,2,3])
print(t1)
print(type(t1))  #<class 'numpy.ndarray'>

2. np.arange(x) The effect of the method and np.array(range(x)) The effect is the same

t2 = np.array(range(10))
print(t2)
print(type(t2))

t3 = np.arange(10)
print(t3)
print(type(t3))

3. Use dtype You can get the data type

t4 = np.arange(4,10,2)
print(t4)
print(type(t4))

print(t4.dtype)

4.numpy Data types in

t5 = np.array(range(1,4),dtype=float)
#  or t5 = np.array(range(1,4),dtype="float")
#  or t5 = np.array(range(1,4),dtype="float32")
print(t5)
print(t5.dtype)

# numpy Medium bool data type 
t6 = np.array([0,1,1,0,1,0,0],dtype="bool")
print(t6)
print(t6.dtype)
print("//")

#  Adjust the data type 
t7 = t5.astype("int8")
print(t5.dtype)
print(t7.dtype)
print("***************************")
t5.dtype = "int8"
print(t5.dtype)

# numpy The decimal in 
t7 = np.array([random.random() for i in range(10)])
print(t7)
print(t7.dtype)

#  Modify floating point scale 
t8 = np.round(t7,2) # Keep two decimal places 
print(t8)

Shape and operation of array

1. a.shape View array shapes

import numpy as np
t1 = np.arange(12)
print(t1)#t1:{ndarray:(12,)}
print(t1.shape)# Output results ：(12,)

t2 = np.array([[1,2,3],[4,5,6]])
print(t2)#t2:{ndarray:(2,3)}
print(t2.shape)# Output results ：(2,3)

t3 = np.array([[[1,2,3],[4,5,6]],[[7,8,9],[10,11,12]]])
print(t3)#t2:{ndarray:(2,2,3)}
print(t3.shape)# Output results ：(2,2,3)

2.a.reshape(3,4) Modify array shape

t4 = np.arange(12)
print(t4)
print(t4.reshape(3,4))
print(t4.reshape(3,5))# Cannot become an array of this shape 

t5 = np.arange(24).reshape(2,3,4)
print(t5)
print("---------------")
# t5.reshape(4,6) There is a return value , Will not change t5 Value 
print(t5.reshape(4,6))
print("---------------")
print(t5)
print("---------------")
#  take t5 Into a one-dimensional array 
print(t5.reshape(24,))
print("---------------")
#  Find out the following two situations 
print(t5.reshape(24,1))
print("---------------")
print(t5.reshape(1,24))
print("#######################")

3.t5.flatten(), take t5 Into a one-dimensional array
Don't know t5 How much data is there , take t5 Into a one-dimensional array

# Law 1 
t5 = np.arange(24).reshape(2,3,4)
print(t5.reshape(t5.shape[0]*t5.shape[1]*t5.shape[2],))
print("#######################")
#  Law two ：
print(t5.flatten())

4. The operation of arrays and numbers

# Operation of array 
t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
print(t6)
print("************************")
print(t6+2)
print("************************")
print(t6-2)
print("************************")
print(t6*2)
print("************************")
print(t6/2)
print("************************")
print(t6/0)

The result is ：

5. Array and array operation
① Situation 1 ：

t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
t7 = np.arange(100,124).reshape(4,6)
print(t7)
print("**************************")
print(t6+t7)
print("**************************")
print(t6-t7)
print("**************************")
print(t6*t7)
print("**************************")
print(t6/t7)

The result is ：

② Situation two ：

t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
t8 = np.arange(0,6)
print(t8)
print(t6)
print(t6-t8)

③ Situation three ：

t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
t9 = np.arange(4).reshape(4,1)
print(t9)
print(t6)
print(t6-t9)

④ Situation four ：

t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
t10 = np.arange(10)
print(t10)
print(t6)
#  Can't calculate 
print(t6-t10)
print("*"*100)

numpy Index and slice of

1. Take one row and one column or multiple rows and multiple columns

import numpy as np
us_file_path = "./dataCSV/data.csv"
t1 = np.loadtxt(us_file_path,delimiter=",",dtype="int",unpack=True)
t2 = np.loadtxt(us_file_path,delimiter=",",dtype="int")
print(t1)
print("*"*100)
print(t2)
print("*"*100)
#  Take row 
print(t2[2])
print("*"*100)
#  Take consecutive lines 
print(t2[2:])
print("*"*100)
#  Take discontinuous lines , Take the first place 2,8,10 That's ok 
print(t2[[2,8,10]])
print("*"*100)
#  Take row  print(t2[ That's ok , Column ])
# print(t2[1,:])
# print(t2[2:,:])
# print(t2[[2,8,10],:])

#  Fetch 
print(t2[:,0]) # Take the data in the first column of all rows 
print("*"*100)
# Take consecutive Columns 
print(t2[:,2:])# Take all column data after the third column of all rows 
print("*"*100)
# Take consecutive Columns 
print(t2[:,[0,2]])## Take the data of the first and third columns of all rows 
print("*"*100)
# Fetching rows and columns , Take the third line , The value of the fourth column 
a = t2[2,3]
print(a)
print(type(a))
print("*"*100)
# Fetching multiple rows and columns , Take the third line to the fifth line , The values of the second to fourth columns ( Take the position where rows and columns intersect )
b = t2[2:5,1:4]
print(b)
print("*"*100)
# Take multiple non adjacent points 
# The result is (0,0) (2,1) (2,3)
c = t2[[0,2,2],[0,1,3]]
print(c)

2. The exchange of ranks

3. Array splicing

①np.vstack(t1,t2) For vertical splicing
②np.hstack(t1,t2) Use horizontal splicing

4. Array transposition

Transpose is a transformation , about numpy For arrays in , Is in the Exchange data diagonally , The purpose is also to process data more conveniently . There are three ways ：
①t.transpose()
②t.swapaxes(1,0)
③t.T

5.numpy Reading data

6.numpy Generate random number

7.numpy Boolean index in

t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
print(t6<10)
print("*"*100)
t6[t6<10]=3
print(t6)
print("*"*100)
print(t6[t6>20])
print("*"*100)
t6[t6>20]=20
print(t6)
print("*"*100)

8. The ternary operation

#  The ternary operation 
# hold t6 Less than equal to 3 Of is replaced by 100, Otherwise replace with 300
t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
print(np.where(t6<=3,100,300))
print("*"*100)
print(t6)
print("*"*100)

## 9. Assign a value to nan
t5 = np.arange(24).reshape(2,3,4)
t6 = t5.reshape(4,6)
t6 = t6.astype(float)
t6[3,3]=np.nan
print(t6)

9.numpy Medium clip

numpy Medium nan and inf

numpy Medium nan Points for attention

Handle ： to nan Assigned mean value

import numpy as np
# to nan Assigned mean value 
def fill_ndarray(t1):
    for i in range(t1.shape[1]):
        temp_col = t1[:,i]
        nan_num = np.count_nonzero(temp_col != temp_col)
        if nan_num != 0:# Not for 0, It indicates that there are nan
            temp_not_nan_col = temp_col[temp_col == temp_col] # The current column is not nan Of array
            # Check that the current is nan The location of , Assign a value that is not nan The average of 
            temp_col[np.isnan(temp_col)] = temp_not_nan_col.mean()
    return t1

if __name__ == '__main__':
    t1 = np.arange(12).reshape((3, 4)).astype("float")
    t1[1, 2:] = np.nan
    print(t1)
    print("*"*100)
    t1 = fill_ndarray(t1)
    print(t1)

numpy Common functions in

numpy Medium copy and view