Question

我正在处理3个数据集，我写了3个不同的函数，每个数据集1个来做一些数据清理和操作。最后，我想在另一个函数中组合所有3个已清理的数据集。

我的逻辑：

import pandas as pd
def function1():
    read in data as df
    df[(df.column1 != "")&(df.column2 != 'MRN')&(df.column3 != "C") ]
    return data1.to_csv() 

def function2():
    read in data as df
    df[(df.column1 != "A")&(df.column2 != 'M')&(df.column3 != " ") ]
    return data2.to_csv() 

def function3():
    read in data as df
    df[(df.column1 != "B")&(df.column2 != 'N')&(df.column3 != " ") ]
    return data3.to_csv() 

def combinedatasets():
    merge (data1, data2, data3)
    return combineddata.to_csv()

现在我将data1，data2和data3作为新文件输出到目录中。有没有将它们临时存储在脚本中，这样就不会输出这3个文件，只输出combineddate.csv？ 如何从我的combineddatasets函数中的前3个函数中调用这些临时数据集data1，data2，data3来组合它们？

如下所示：

import pandas as pd
def function1():
    read in data as df
    df[(df.column1 != "")&(df.column2 != 'MRN')&(df.column3 != "C") ]
    return temporary data1 without outputting it

def function2():
    read in data as df
    df[(df.column1 != "A")&(df.column2 != 'M')&(df.column3 != " ") ]
    return temporary data2 without outputting it

def function3():
    read in data as df
    df[(df.column1 != "B")&(df.column2 != 'N')&(df.column3 != " ") ]
    return temporary data3 without outputting it

def combinedatasets():
    calling temporary data1,2,3 and 
    merge (data1, data2, data3)
    return pd.to_csv('combineddata.csv') #output as a csv file

因此只有'combineddata.csv'会输出到该文件夹。

Answer 1

简单地将一个对象分配给函数调用，因为函数返回一个数据帧：

def myfunction():
    data = pd.read_csv('Input.csv')
    # process dataframe...
    return data

def combinedatasets():
    df = myfunction()

或同时分配：

def combinedatasets():
    data1, data2, data3 = function1(), function2(), function3()

但是，请避免在您的环境中使用类似结构的多个数据帧，并将数据帧保存到一个列表中，然后您可以合并或追加这些列表：

def combinedatasets():
    dfList = [function1(), function2(), function3()]

    # MERGE/COLUMN BIND
    combinedf = pd.concat(dfList, axis=1, join_axes=[dfList[0].index])
    combinedf.to_csv('CombinedWideData.csv')

    # APPEND/ROW BIND
    combinedf = pd.concat(dfList)
    combinedf.to_csv('CombinedLongData.csv')

使用来自不同函数python pandas的多个返回数据集

1 个答案: