Hands on deep learning pytorch version exercise answer - 2.2 preliminary knowledge / data preprocessing

I am here jupyter notebook Completed exercises , It takes a little more time. After all, it's new , Like judging numbers 、 Judge whether the value is empty 、 Delete the specified columns , Give a reference to friends who read this book newly . First post the overall operation results , Code after

Here's the code section ：

import os
os.makedirs(os.path.join('..','practice'),exist_ok=True)
practice_file=os.path.join('..','practice','student_scores.csv')
with open(practice_file,'w') as f:
    f.write('stu_num,stu_name,stu_course,stu_score\n')
    f.write('1,Lily,English,100\n')
    f.write('1,Lily,Physics,80\n')
    f.write('2,NA,Computer Science,90\n')
    f.write('2,NA,Database,88\n')
    f.write('NA,John,Math,99\n')
    f.write('4,Lisa,NA,100\n')
    f.write('5,NA,French,50\n')
    f.write('6,GOGO,NA,10\n')

# Delete the column with the most missing values 
import pandas as pd
practice_data=pd.read_csv(practice_file)
print(practice_data)

# Take the horizontal and longitudinal length 
column_len,row_len=len(practice_data.iloc[0,:]),len(practice_data.iloc[:,0])
# Count each column nan Number of 
nan_sum=[0 for i in range(column_len)]
print(" Column length ：{}, Line length ：{}, Statistics nan Array of ：{} ".format(column_len,row_len,nan_sum))
import math
import numbers
for i in range(column_len):
    for j in range(row_len):
        #NaN There are two kinds of ： Numbers and strings , This should be judged separately 
        # Note that the number one j+1 That's ok 、 The first i+1 Column 
        if isinstance(practice_data.iloc[j,i],numbers.Number) == True:
            if math.isnan(practice_data.iloc[j,i]) == True:
                nan_sum[i] += 1
        else:
            if practice_data.iloc[j,i]=='NaN':
                nan_sum[i] += 1        
max_index=[]# The reason for getting the list is that there may be multiple columns with the most nan
most_nan=max(nan_sum)
for i,j in enumerate(nan_sum):
    if j == most_nan:
        max_index.append(practice_data.columns.values[i])
print(" The column with the most missing values is {}".format(max_index))

print(" The column with the most missing values is not deleted ：")
print(practice_data)

# Start deleting the column with the most missing values 
# This drop Delete the name of the column , Remember to update 
practice_data = practice_data.drop(max_index,axis=1)
print(" After deleting the column with the most missing values ")
print(practice_data)

# The preprocessed data set is transformed into tensor
practice_inputs = practice_data.iloc[:,:]
practice_inputs = pd.get_dummies(practice_inputs,dummy_na=True)
print(practice_inputs)
import torch
Z = torch.tensor(practice_inputs.values)
print(Z)