编辑python pandas过滤器中的数据并将其应用于原始数据框

Question

I am trying to figure out how to filter data in pandas then assign a value to all of the rows in a column for the items that meet the filter criteria and have it affect the original data frame. Here is the closest attempt I have so far but it is throwing a lot of informational warnings:

    import pandas as pd
    df = pd.read_csv('http://www.sharecsv.com/dl/9096d32f98aa0ac671a1cca16fa43be8/SalesJan2009.csv')
    df['Zone'] = ''
    zone1 = df[(df['Latitude'] > 0) & (df['Latitude'] > 0)]
    zone2 = df[(df['Latitude'] < 0) & (df['Latitude'] > 0)]
    zone3 = df[(df['Latitude'] > 0) & (df['Latitude'] < 0)]
    zone4 = df[(df['Latitude'] < 0) & (df['Latitude'] < 0)]
    zone1[['Zone']] = zone1[['Zone']] = 1
    zone2[['Zone']] = zone1[['Zone']] = 2
    zone3[['Zone']] = zone1[['Zone']] = 3
    zone4[['Zone']] = zone1[['Zone']] = 4
    df

This does not affect the original data frame at all but it is setting the values in the filtered subsets.

I am assuming that I may need to filter out everything that meets each of my filters and remove it from the original and then concatenate the changes back onto the original??

This is a random dataset to illustrate what I am looking to do but my actual dataset has data that doesn't meet any filter criteria and I need to maintain those as unknown as well because I am not consuming all rows as I would be with this example.

I am trying to avoid having to loop over every row and check criteria against every row so if anyone knows how I can accomplish this I would be super grateful!

Answer 1

IIUC, are you trying to do something like this:

zone1 = (df['Latitude'] > 0) & (df['Longitude'] > 0)
zone2 = (df['Latitude'] < 0) & (df['Longitude'] > 0)
zone3 = (df['Latitude'] > 0) & (df['Longitude'] < 0)
zone4 = (df['Latitude'] < 0) & (df['Longitude'] < 0)

df['Zone'] = np.select([zone1,zone2,zone3,zone3],['Zone 1','Zone 2', 'Zone 3','Zone 4'])

Output:

  Transaction_date   Product Price Payment_Type               Name  \
0      1/2/09 6:17  Product1  1200   Mastercard           carolina   
1      1/2/09 4:53  Product1  1200         Visa             Betina   
2     1/2/09 13:08  Product1  1200   Mastercard  Federica e Andrea   
3     1/3/09 14:44  Product1  1200         Visa              Gouya   
4     1/4/09 12:56  Product2  3600         Visa            Gerd W    

                           City     State         Country Account_Created  \
0                      Basildon   England  United Kingdom     1/2/09 6:00   
1  Parkville                           MO   United States     1/2/09 4:42   
2  Astoria                             OR   United States    1/1/09 16:21   
3                        Echuca  Victoria       Australia   9/25/05 21:13   
4  Cahaba Heights                      AL   United States  11/15/08 15:47   

     Last_Login   Latitude   Longitude    Zone  
0   1/2/09 6:08  51.500000   -1.116667  Zone 3  
1   1/2/09 7:49  39.195000  -94.681940  Zone 3  
2  1/3/09 12:32  46.188060 -123.830000  Zone 3  
3  1/3/09 14:22 -36.133333  144.750000  Zone 2  
4  1/4/09 12:45  33.520560  -86.802500  Zone 3  

Answer 2

You missed that both conditions are checking for Latitude and you should check out .loc so you learn how to change values in parts of the dataframe the right way.

import pandas as pd
df = pd.read_csv('http://www.sharecsv.com/dl/9096d32f98aa0ac671a1cca16fa43be8/SalesJan2009.csv')
df['Zone'] = ''
zone1 = (df['Latitude'] > 0) & (df['Longitude'] > 0)
zone2 = (df['Latitude'] < 0) & (df['Longitude'] > 0)
zone3 = (df['Latitude'] > 0) & (df['Longitude'] < 0)
zone4 = (df['Latitude'] < 0) & (df['Longitude'] < 0)
df.loc[zone1, 'Zone'] = 1
df.loc[zone2, 'Zone'] = 2
df.loc[zone3, 'Zone'] = 3
df.loc[zone4, 'Zone'] = 4
df