PANDAS - 如何根据另一个 dataframe 中的查找表替换 dataframe 中的值

Question

I have two csv files:

live_file.csv

Supplier SKU, Manufacturer SKU, Price
ABCD, 900000, 10
EFGH, 800000, 10

old_file.csv

Supplier SKU, Manufacturer SKU, Price
ABCD, 91234, 10
EFGHX, 85332, 10

I want to find the same values in the column Supplier SKU column, when I find matching vaues I want to take the Manufacturer SKU value from old_file.csv and put it in the live_file.csv, so my result will be:

Supplier SKU, Manufacturer SKU, Price
ABCD, 91234, 10
EFGH, 800000, 10

This is what i tried:

import pandas as pd

live_file  = pd.read_csv("live.csv")
old_file   = pd.read_csv("old.csv") 

old_file = old_file.set_index('Supplier SKU')['Manufacturer SKU'].dropna()
live_file['Manufacturer SKU'] = live_file['Supplier SKU'].replace(old_file)

live_file.to_csv(r'final.csv')

But this does not work, the end file is the same as the live file at the beginning, any help?

Answer 1

Set index on 'Supplier SKU' using set_index , then call update on new livefile.

import pandas as pd

df1 = pd.DataFrame({
    'Supplier SKU': ['ABCD','EFGH'],
    'Manufacturer SKU': [900000, 800000],
    'Price': [10, 10]
}).set_index('Supplier SKU')
df2 = pd.DataFrame({
    'Supplier SKU': ['ABCD','EFGHX'],
    'Manufacturer SKU': [91234, 85332],
    'Price': [10, 10]
}).set_index('Supplier SKU')

df1.update(df2)
print(df1)

result:

To prevent 'Price' also being updated, you can just drop Price in df2:

df1.update(df2.drop(columns='Price'))

PS: call df1.reset_index() to make 'Supplier SKU' into ordinary column

Answer 2

You can basically do a left-merge join of the two files on the column Supplier SKU and then keep the value of column Manufacturer SKU from old_file when the merge matches, otherwise keep the value from live_file

live_file["Manufacturer SKU"] =  pd.merge(live_file[["Supplier SKU", "Manufacturer SKU"]], 
         old_file[["Supplier SKU", "Manufacturer SKU"]],
         how="left", 
         on="Supplier SKU",
         suffixes=(None, "__right"),
         indicator="merge_flag")\
         .apply(lambda row: row["Manufacturer SKU"] 
                if row["merge_flag"] == "left_only" 
                else row["Manufacturer SKU__right"], axis=1)