Pandas_VBA_数据清洗对比

VBA处理数据与Python Pandas处理数据案例比较

Author : Collin_PXY

需求：

现有一个 csv文件，包含'CNUM'和'COMPANY'两列，数据里包含空行，且有内容重复的行数据。

要求：

1）去掉空行；

2）重复行数据只保留一行有效数据；

3）修改'COMPANY'列的名称为'Company_New‘；

4）并在其后增加六列，分别为'C_col','D_col','E_col','F_col','G_col','H_col'。

一，使用 Python Pandas来处理：

import pandas as pd
import numpy as np
from pandas import DataFrame,Series

def deal_with_data(filepath,newpath):
    file_obj=open(filepath)
    df=pd.read_csv(file_obj)    # 读取csv文件，创建 DataFrame
    df=df.reindex(columns=['CNUM','COMPANY','C_col','D_col','E_col','F_col','G_col','H_col'],fill_value=None)   # 重新指定列索引
    df.rename(columns={'COMPANY':'Company_New'}, inplace = True)  # 修改列名
    df=df.dropna(axis=0,how='all')                 # 去除 NAN 即文件中的空行
    df['CNUM'] = df['CNUM'].astype('int32')        # 将 CNUM 列的数据类型指定为 int32
    df = df.drop_duplicates(subset=['CNUM', 'Company_New'], keep='first')  # 去除重复行
    df.to_csv(newpath,index=False,encoding='GBK')
    file_obj.close()

if __name__=='__main__':
    file_path=r'C:\Users\12078\Desktop\python\CNUM_COMPANY.csv'
    file_save_path=r'C:\Users\12078\Desktop\python\CNUM_COMPANY_OUTPUT.csv'
    deal_with_data(file_path,file_save_path)

二，使用 VBA来处理：

Option Base 1
Option Explicit

Sub main()
	On Error GoTo error_handling
	Dim wb                  As Workbook
	Dim wb_out              As Workbook
	Dim sht                 As Worksheet
	Dim sht_out             As Worksheet
	Dim rng                 As Range
	Dim usedrows            As Byte
	Dim usedrows_out        As Byte
	Dim dict_cnum_company   As Object
	Dim str_file_path       As String
        Dim str_new_file_path   As String

        'assign values to variables:
       str_file_path = "C:\Users\12078\Desktop\Python\CNUM_COMPANY.csv"
       str_new_file_path = "C:\Users\12078\Desktop\Python\CNUM_COMPANY_OUTPUT.csv"

	Set wb = checkAndAttachWorkbook(str_file_path)
	Set sht = wb.Worksheets("CNUM_COMPANY")
	Set wb_out = Workbooks.Add
	wb_out.SaveAs str_new_file_path, xlCSV  'create a csv file
	Set sht_out = wb_out.Worksheets("CNUM_COMPANY_OUTPUT")

	Set dict_cnum_company = CreateObject("Scripting.Dictionary")
	usedrows = WorksheetFunction.Max(getLastValidRow(sht, "A"), getLastValidRow(sht, "B"))

	'rename the header 'COMPANY' to 'Company_New',remove blank & duplicate lines/rows.
	Dim cnum_company As String
	cnum_company = ""
	For Each rng In sht.Range("A1", "A" & usedrows)
	    If VBA.Trim(rng.Offset(0, 1).Value) = "COMPANY" Then
	        rng.Offset(0, 1).Value = "Company_New"
	    End If
	    cnum_company = rng.Value & "-" & rng.Offset(0, 1).Value
	    If VBA.Trim(cnum_company) <> "-" And Not dict_cnum_company.Exists(rng.Value & "-" & rng.Offset(0, 1).Value) Then
	        dict_cnum_company.Add rng.Value & "-" & rng.Offset(0, 1).Value, ""
	    End If
	Next rng

	'loop the keys of dict split the keyes by '-' into cnum array and company array.
	Dim index_dict As Byte
	Dim arr_cnum()
	Dim arr_Company()
	For index_dict = 0 To UBound(dict_cnum_company.keys)
	    ReDim Preserve arr_cnum(1 To UBound(dict_cnum_company.keys) + 1)
	    ReDim Preserve arr_Company(1 To UBound(dict_cnum_company.keys) + 1)
	    arr_cnum(index_dict + 1) = Split(dict_cnum_company.keys()(index_dict), "-")(0)
	    arr_Company(index_dict + 1) = Split(dict_cnum_company.keys()(index_dict), "-")(1)
	    Debug.Print index_dict
	Next

	'assigns the value of the arrays to the celles.
	sht_out.Range("A1", "A" & UBound(arr_cnum)) = Application.WorksheetFunction.Transpose(arr_cnum)
	sht_out.Range("B1", "B" & UBound(arr_Company)) = Application.WorksheetFunction.Transpose(arr_Company)

	'add 6 columns to output csv file:
	Dim arr_columns() As Variant
	arr_columns = Array("C_col", "D_col", "E_col", "F_col", "G_col", "H_col")   '
	sht_out.Range("C1:H1") = arr_columns
	Call checkAndCloseWorkbook(str_file_path, False)
	Call checkAndCloseWorkbook(str_new_file_path, True)

Exit Sub
error_handling:
   Call checkAndCloseWorkbook(str_file_path, False)
   Call checkAndCloseWorkbook(str_new_file_path, False)
End Sub

' 辅助函数：
'Get last row of Column N in a Worksheet
Function getLastValidRow(in_ws As Worksheet, in_col As String)
    getLastValidRow = in_ws.Cells(in_ws.Rows.count, in_col).End(xlUp).Row
End Function

Function checkAndAttachWorkbook(in_wb_path As String) As Workbook
    Dim wb As Workbook
    Dim mywb As String
    mywb = in_wb_path

   For Each wb In Workbooks
        If LCase(wb.FullName) = LCase(mywb) Then
            Set checkAndAttachWorkbook = wb
            Exit Function
        End If
    Next

   Set wb = Workbooks.Open(in_wb_path, UpdateLinks:=0)
   Set checkAndAttachWorkbook = wb

End Function

Function checkAndCloseWorkbook(in_wb_path As String, in_saved As Boolean)
    Dim wb As Workbook
    Dim mywb As String
    mywb = in_wb_path
    For Each wb In Workbooks
        If LCase(wb.FullName) = LCase(mywb) Then
            wb.Close savechanges:=in_saved
            Exit Function
        End If
    Next
End Function

三，输出结果：

两种方法输出结果相同：

四，比较总结：

Python pandas 内置了大量处理数据的方法，我们不需要重复造轮子，用起来很方便，代码简洁的多。

Excel VBA 处理这个需求，使用了 数组，字典等数据结构（实际需求中，数据量往往很大，所以一些地方没有直接使用遍历单元格的方法），以及处理字符串，数组和字典的很多方法，对文件的操作也很复杂，一旦出错，调试起来比python也较困难，代码已经尽量优化，但还是远比 Python要多。