如何使用 python 基于键值合并数据帧或 CSV 文件？

Question

I have two csv files, first one (master file) contain Keys and its values and the second one (daily file) contain key and some other columns.

Example file (master file)

  Key  value
   A     1
   B     2
   c     3

Example file (daily file)

 Name  Key  date
 Red   A    dd/mm/yy
 Blue  B    dd/mm/yy
 Pink  C    dd/mm/yy

The outcome file I need looks like this:

 Name   Key   value   date
 Pink    C      3     dd/mm/yy
 Blue    B      2     dd/mm/yy
 Red     A      1     dd/mm/yy

I've tried using dataframe and creating dicts from external file or dataframe but not idea how can I do the lookup base on key and obtain its value.

Answer 1

Use this code, may be it will help you and you will get your desire output

# import Library
import pandas as pd 

# Create Dataframe as like as imported CSV file for your (master file) and (daily file)
masterfile = {'key':['A','B','C'], 'value':[1,2,3]}
dailyfile = {'Name':['Red','Blue','Pink'],'key':['A','B','C'], 'date':['dd/mm/yy','dd/mm/yy','dd/mm/yy']}

masterfil = pd.DataFrame(data=masterfile, index=None)
dailyfile = pd.DataFrame(data=dailyfile, index=None)

# Change the order of dataframe(descending order)
df_masterfil = masterfil.sort_values(by='key', ascending=False)
df_dailyfile = dailyfile.sort_values(by='key', ascending=False)

# merge the both dataframe or csv file
df=df_dailyfile.merge(df_masterfil)
df=df[['key', 'value', 'date']]

dailyfile['Name']

# As like your dataset
result= pd.concat([dailyfile['Name'], df1], axis=1)
result

Answer 2

i think the most basic answer to your question is given by w3schools .

How you first put the content of your files into dicts is a different story that starts by reading the file . I think i would choose the readline() option and split() ting the gotten strings into key and value pairs for the dict()

Answer 3

If your csv files as you've described, you can do this very easily with Pandas to merge records from both files with matching Key values:

import pandas as pd
df1 = pd.read_csv('master.csv')
df2 = pd.read_csv('daily.csv')
df3 = df2.merge(df1, left_on='Key', right_on='Key')

This gives you a merged dataframe for all data with matching Key s:

   Name Key      date  value
0   Red   A  dd/mm/yy      1
1  Blue   B  dd/mm/yy      2
2  Pink   C  dd/mm/yy      3

If you want the columns in the order in your question you can just add

df3 = df3[['Name', 'Key', 'value', 'date']]