Unpleasant error typeerror: cannot perform 'ROR_‘ with a dtyped [float64] array and scalar of type [bool]

Hello everyone , I'm an old watch , This series will record some problems and solutions I encounter in programming , Will directly take the error report encountered as the title , In this way, it is more convenient for other learners to search for problem solutions in the browser , You are also welcome to comment on 、 More communication in the message area , Talk about your understanding or problems .

such as Unpleasant –requests.exceptions.ProxyError Is a good demonstration .

Today, when processing a data , Executed a line of code , As a result, there was a big problem ！！！

#  Simple data processing , Remove recommended hot search and top hot search 
wb_hot_data.drop(wb_hot_data[wb_hot_data['wb_rank']=='•' | pd.isna(wb_hot_data['wb_rank'])].index)

Look at this pile of errors , I guess the code is also wronged ！！！

Let's talk directly about how to solve ：

#  Simple data processing , Remove recommended hot search and top hot search 
wb_hot_data.drop(wb_hot_data[(wb_hot_data['wb_rank']=='•') | pd.isna(wb_hot_data['wb_rank'])].index)

Maybe you don't see what has changed , In fact, it's for wb_hot_data['wb_rank']=='•' Added a bracket .

Next, let's talk about how this error occurs with example data ？

import pandas as pd
import numpy as np
 
a=np.array([['a',2,3],['b',5,6],[np.NaN,8,9],['a',4,7]])
df1=pd.DataFrame(a,columns=list('ABC'))

print(df1)

''' A B C 0 a 2 3 1 b 5 6 2 nan 8 9 3 a 4 7 '''

I want to delete A The value in the column is ’a’ perhaps None The line of :

df1.drop(df1[df1['A']=='a' | pd.isna(df1['A'])].index)

There seems to be nothing wrong with writing like this , But when running, there will be annoying errors in the title .

So Google checked , eureka An article with the same mistake , Bloggers in the comment area put parentheses around an expression , It's a priority issue .

See the answer , I understand , If you don't have a solid foundation, you will suffer losses .

df1.drop((df1[df1['A']=='a') | pd.isna(df1['A'])].index)
#  Just add a bracket 

''' A B C 1 b 5 6 2 nan 8 9 '''

We'll find here nan Not deleted ？ Because the initialization is this nan Has been automatically converted to a string , So it can't be recognized as nan Be deleted , Let's talk about this in other articles , You can also leave a message to say your views .

Add another case , When A Column as numeric data , See what happens ？

import pandas as pd
import numpy as np
 
# a=np.array([['a',2,3],['b',5,6],[np.NaN,8,9],['a',4,7]])
a=np.array([[1,2,3],[2,5,6],[np.NaN,8,9],[2,4,7]])
df1=pd.DataFrame(a,columns=list('ABC'))

print(df1)

''' A B C 0 1.0 2.0 3.0 1 2.0 5.0 6.0 2 NaN 8.0 9.0 3 2.0 4.0 7.0 '''

#  I want to delete A The value in the column is 2 perhaps None The line of 
df1.drop(df1[df1['A']==2 | pd.isna(df1['A'])].index)

''' A B C 1 2.0 5.0 6.0 2 NaN 8.0 9.0 3 2.0 4.0 7.0 '''

At this time, we found out , No report error , But the result is obviously problematic , Why is that ？

This is because numeric types can be directly associated with bool Type data , So there was no error ; The result is wrong because by operator priority , to 2 | pd.isna(df1['A']), What comes out is true（2 Is a definite number , stay bool The type is true,true Comparing with the numerical value will turn into 1）, And then df1[‘A’] Conduct == operation , You'll find only the first line 1 Is satisfied (1==ture), So the first line will be deleted .
So it's still operator priority , Or add () You can solve the problem .

df1.drop(df1[(df1['A']==2) | pd.isna(df1['A'])].index)
''' A B C 0 1.0 2.0 3.0 '''
#  Why can we put nan Value delete ？ You can talk about 

So the problem we have today is caused by operator priority String type data and bool Type data cannot be | operation .

The last attached Python List of operator priority and specificity .

Reference resources

[1] stackoverflow Answer relevant questions
https://stackoverflow.com/questions/20333435/pandas-comparison-raises-typeerror-cannot-compare-a-dtyped-float64-array-with
[2] Python List of operator priority and specificity
http://c.biancheng.net/view/2190.html