Python实现Excel办公自动化
Python实现Excel办公自动化
在工作中难免遇到统计表格，汇总表格，数据查重等。可根据自己的需求进行更改，代码书写的初衷主要是想解法双手，摆脱重复机械的ctrl+C，ctrl+V，完成数据的批量替换，查找不同表格的相同位置，用代码的1秒钟代替你一小时的复制粘贴。
1 开发环境：
Python3.6及以上
pandas，matplotlib：数据分析及可视化
xlwings：配合系统的office，完成功能扩展
2 具体功能及实现代码
- 批量新建Excel文件
import xlwings as xw
app=xw.App(visible=True,add_book=False)
for grade in ["一年级","二年级","三年级","四年级"]:
workbook=app.books.add()
workbook.save(f"./信息统计-{grade}.xlsx")
- 批量打开Excel文件
import os
import xlwings as xw
app=xw.App(visible=True,add_book=False)
for file in os.listdir("."):
if file.endswith(".xlsx"):
app.books.open(file)
将会打开".xlsx"结尾的文件：
- 批量重命名Excel工作表
import xlwings as xw
app=xw.App(visible=False,add_book=False)
workbook=app.books.open("信息统计-一年级.xlsx")
for sheet in workbook.sheets:
sheet.name=sheet.name.replace("Sheet","成绩")
workbook.save()
app.quit()
初始：
代码运行后：
- 批量合并Excel文件
import pandas as pd
import os
data_list=[]
for fname in os.listdir("."):
if fname.startswith("信息统计-") and fname.endswith(".xlsx"):
data_list.append(pd.read_excel(fname))
data_all=pd.concat(data_list)
data_all.to_excel("信息统计总表.xlsx",index=False)
表一：
表二：
运行效果：
功能完善中…
3 具体需求中的应用：
txt文档替换内容，数学建模初步筛选数据使用
def updateFile(file,old_str,new_str):
file_data = ""
with open(file, "r", encoding="utf-8") as f:
for line in f:
if old_str in line:
line = line.replace(old_str,new_str)
file_data += line
with open(file,"w",encoding="utf-8") as f:
f.write(file_data)
def updateOneLine(file):
file_data = ""
with open(file, 'r', encoding="utf-8") as r:
lines = r.readlines()
with open(file, "w", encoding="utf-8") as f:
for l in lines:
if 'DecaRangeRTLS' not in l:
f.write(l)
else:
f.write("time id distance real_distance serial_number number \n")
for i in range(1,325):
flie_name=str(i)+".异常.txt"
print(flie_name)
updateOneLine(flie_name)
updateFile(flie_name, "T:", "")
updateFile(flie_name, ":RR:0:", " ")
updateFile(flie_name, ":", " ")
Excel表格合成（汇总数据使用）汇总多人的表格数据，快速合并到对应的表格中
import os # 导入模块
import pandas as pd
def combine_Excel(dir):
file_name_li = os.listdir(dir)
# 遍历出每个文件名
for file_name in file_name_li:
# 将文件夹绝对路径 与 文件名进行拼接
file_path_li = os.path.join(dir, file_name)
print(file_path_li)
# 定义文件名集合
# all_file_name = set()
all_file_name = []
# 定义数据列表
all_data_li = []
# 遍历出每个文件名
for file_name in file_name_li:
# 将文件夹绝对路径 与 文件名进行拼接
file_path_li = os.path.join(dir, file_name)
# 读取 excel 表格数据
all_data = pd.read_excel(file_path_li, sheet_name=None)
# 将数据添加到数据列表中
all_data_li.append(all_data)
# 将工作表名添加到文件夹集合中
for name in all_data:
all_file_name.append(name)
# print(all_data_li)
# all_file_name.sort()
new_all_file_name = []
for x in all_file_name:
if x not in new_all_file_name:
new_all_file_name.append(x)
print(new_all_file_name)
# 创建工作簿
writer = pd.ExcelWriter("all_data.xlsx")
# 遍历每个工作表名
for sheet_name in all_file_name:
data_li = []
# 遍历数据
for data in all_data_li:
# 获取同名数据并添加到data_li中
try:
n_rows = data_li.append(data[sheet_name])
except:
print("此表格无此表！")
continue
# 将同名数据进行拼接
group_data = pd.concat(data_li)
# 保存到writer工作簿中，并指定工作表名为sheet_name
group_data.to_excel(writer, sheet_name=sheet_name)
# 千万莫忘记，保存工作簿
writer.save()
if __name__ == '__main__':
dir = r'C:\Users\dell\Desktop\Test_EXCel\Excel'
combine_Excel(dir)
pass
Excel 统一时间格式，主要是完成表格中的数据替换
# coding=utf-8
import xlrd
import codecs
import datetime
def clear_date_xlsx(file,sheet,num):
# 加载Excel数据，处理数据
data = xlrd.open_workbook(file) # 读取工作表
table = data.sheet_by_name(sheet) # 读取当前sheet表对象
rows = table.nrows # 获取行数
print('一共有{}行数据，开始清洗数据'.format(rows))
for i in range(1, rows):
MyTime = str(table.row_values(i)[num])
today=" "
if "44" in MyTime:
delta=datetime.timedelta(days=float(MyTime))
today = datetime.datetime.strptime('1899-12-30', '%Y-%m-%d') + delta
today=datetime.datetime.strftime(today, '%Y.%m.%d')
if ".0" in today:
today = today.replace(".0", ".")
elif "/" in MyTime:
today = datetime.datetime.strptime(MyTime, '%m/%d/%Y')
today = datetime.datetime.strftime(today, '%Y.%m.%d')
if ".0" in today:
today = today.replace(".0", ".")
elif "日"in MyTime:
today = datetime.datetime.strptime(MyTime, '%Y年%m月%d日')
today = datetime.datetime.strftime(today, '%Y.%m.%d')
if ".0" in today:
today = today.replace(".0", ".")
elif "年" in MyTime:
today = datetime.datetime.strptime(MyTime, '%Y年%m月')
today = datetime.datetime.strftime(today, '%Y.%m')
if ".0" in today:
today = today.replace(".0", ".")
elif "，" in MyTime:
today = MyTime.replace("，", ".")
else:
today=MyTime
# MyTime = MyTime.replace("年", ".").replace("月","").replace("，", ".")
output = ('{}\n').format(today)
f = codecs.open('清洗后的数据.xls', 'a+')
f.write(output)
f.close()
if __name__ == '__main__':
clear_date_xlsx('all_data.xlsx','sheet',6)
Excel 两个表格找相同数据，避免重复统计
import openpyxl
from openpyxl.styles import PatternFill
import re
import traceback
changeCells = 0
def Find_Same(file1,num1,col1,file2,num2,col2):
# load the file(*.xlsx)
wb1 = openpyxl.load_workbook(file1)
wb2 = openpyxl.load_workbook(file2)
# ! deal with one sheet
ws1 = wb1.worksheets[num1]
ws2 = wb2.worksheets[num2]
global changeCells
# get rows and columns of file
rows1 = ws1.max_row
rows2 = ws2.max_row
all_file1_data = []
file1_cells = 0
same_Cells=0
for row in range(2, rows1+1):
content = ws1.cell(row=row, column=col1).value
if (content != None):
if (type(content) == str):
if "-" in content:
content= content.replace("-", " ")
all_file1_data.append(content.lower())
file1_cells += 1
for row in range(2, rows2+1):
content = ws2.cell(row=row, column=col2).value
Color = ['ffeb9c', '9c6500'] # 黄
fille = PatternFill('solid', fgColor=Color[0])
if (content != None):
if (type(content) == str):
if "-" in content:
content = content.replace("-", " ")
for file1_data in all_file1_data:
if content.lower() in file1_data:
same_Cells += 1
ws2.cell(row=row, column=col2, value=content).fill = fille
ws2.cell(row=row, column=col2).value = content.replace(content, content+" "+str(all_file1_data.index(file1_data)))
print('file1_cells', file1_cells)
print('same_Cells', same_Cells)
wb2.save(file2)
if __name__ == "__main__":
Find_Same('test.xlsx',1,4,'test1.xlsx',1,4)
课代表查出勤,10班是我班学生的名单，遍历搜索出勤名单中未出现的人，出现的表蓝，未出现直接print（他的名字），老师再也不用担心我的学习了，哈哈哈哈哈
import openpyxl
from openpyxl.styles import PatternFill
def Find_Same(file1,num1,col1,file2,num2,col2):
# Find_Same('10班.xlsx',1,4,'test1.xlsx',1,4)
wb1 = openpyxl.load_workbook(file1)
wb2 = openpyxl.load_workbook(file2)
# ! deal with one sheet
ws1 = wb1.worksheets[num1]
ws2 = wb2.worksheets[num2]
# get rows and columns of file
rows1 = ws1.max_row
rows2 = ws2.max_row
find_name_flag = 0
for row in range(2, rows1 + 1):
content = ws1.cell(row=row, column=col1).value
Color = ['ffeb9c', 'A4BBEE'] # 黄
fille = PatternFill('solid', fgColor=Color[1])
if (content != None):
if (type(content) == str):
for row in range(2, rows2 + 1):
attendance_data = ws2.cell(row=row, column=col2).value
if (attendance_data != None):
if (type(attendance_data) == str):
if content in attendance_data:
ws2.cell(row=row, column=1, value=attendance_data).fill = fille
find_name_flag = 1
if row == rows2 and find_name_flag==0:
print("未出勤人员——"+content)
if find_name_flag == 1:
find_name_flag=0
break
wb2.save(file2)
if __name__ == "__main__":
Find_Same('10班.xlsx',0,3,'出勤.xlsx',0,1)
名单核对系统源码
采用pyqt开发显示页面，可快速统计网课人员的出勤情况
主文件：
from PyQt5 import QtWidgets
from PyQt5.QtWidgets import QMessageBox
from PyQt5.QtCore import QTimer
from Ui_show import Ui_Form
from PyQt5.QtWidgets import *
from PyQt5 import QtWidgets
import sys
import os
import openpyxl
from openpyxl.styles import PatternFill
class Pyqt5_Serial(QtWidgets.QWidget, Ui_Form):
def __init__(self):
super(Pyqt5_Serial, self).__init__()
self.setupUi(self)
self.init()
self.setWindowTitle("名单核对系统")
self.file1= ""
self.num1= 0
self.col1 = 0
self.file2 = ""
self.num2 = 0
self.col2 = 0
def init(self):
self.class_Button.clicked.connect(self.open_class_file)
self.work_Button.clicked.connect(self.open_work_file)
self.run_Button.clicked.connect(self.check_excel)
self.class_row.textChanged.connect(self.class_textchanged)
self.woek_row.textChanged.connect(self.work_textchanged)
def open_class_file(self):
fileName, fileType = QtWidgets.QFileDialog.getOpenFileName(self, "选取班级名单", os.getcwd(),
"All Files(*);;Text Files(*.txt)")
if fileName != "":
self.class_lineEdit.setText(fileName)
self.file1 = fileName
def open_work_file(self):
fileName, fileType = QtWidgets.QFileDialog.getOpenFileName(self, "选取出勤名单", os.getcwd(),
"All Files(*);;Text Files(*.txt)")
if fileName != "":
self.work_lineEdit.setText(fileName)
self.file2 = fileName
def class_textchanged(self,text):
self.col1 = int(text)
# print(self.col1)
def work_textchanged(self,text):
self.col2 = int(text)
# print(self.col2)
def check_excel(self):
self.num1 = 0
self.num2 = 0
res_1=QMessageBox.question(self, 'Is it right?',"名单第"+str(self.col1)+"列"+"与"+"出勤第"+str(self.col2)+"列", QMessageBox.Yes | QMessageBox.No, QMessageBox.Yes)
if res_1==QMessageBox.Yes:
try:
self.Find_Same(self.file1, self.num1, self.col1, self.file2, self.num2, self.col2)
QMessageBox.information(self, "Finish!", "核对完成,请查看核对名单.xlsx")
except:
QMessageBox.critical(self, "Error！", "您输入的信息有误！")
else:
pass
def Find_Same(self,file1, num1, col1, file2, num2, col2):
current_path = os.path.abspath(self.file2)
father_path =os.path.dirname(current_path) + os.path.sep
# Find_Same('10班.xlsx',1,4,'test1.xlsx',1,4)
wb1 = openpyxl.load_workbook(file1)
wb2 = openpyxl.load_workbook(file2)
# ! deal with one sheet
ws1 = wb1.worksheets[num1]
ws2 = wb2.worksheets[num2]
# get rows and columns of file
rows1 = ws1.max_row
rows2 = ws2.max_row
find_name_flag = 0
for row in range(2, rows1 + 1):
content = ws1.cell(row=row, column=col1).value
Color = ['F1F84F', 'A4BBEE'] # 颜色
fille = PatternFill('solid', fgColor=Color[1])
if (content != None):
if (type(content) == str):
for row in range(2, rows2 + 1):
attendance_data = ws2.cell(row=row, column=col2).value
if (attendance_data != None):
if (type(attendance_data) == str):
if content in attendance_data:
ws2.cell(row=row, column=1, value=attendance_data).fill = fille
find_name_flag = 1
if row == rows2 and find_name_flag == 0:
self.Show.insertPlainText("未出勤人员——" + content+"\n")
print("未出勤人员——" + content)
if find_name_flag == 1:
find_name_flag = 0
break
wb2.save(father_path+"核对名单.xlsx")
if __name__ == '__main__':
app = QtWidgets.QApplication(sys.argv)
myshow = Pyqt5_Serial()
myshow.show()
sys.exit(app.exec_())
显示页面Ui_show.py
# -*- coding: utf-8 -*-
# Form implementation generated from reading ui file 'test.ui'
#
# Created by: PyQt5 UI code generator 5.15.4
#
# WARNING: Any manual changes made to this file will be lost when pyuic5 is
# run again. Do not edit this file unless you know what you are doing.
from PyQt5 import QtCore, QtGui, QtWidgets
class Ui_Form(object):
def setupUi(self, Form):
Form.setObjectName("Form")
Form.resize(631, 246)
Form.setStyleSheet("background-color: rgb(252, 208, 255);")
self.verticalLayoutWidget = QtWidgets.QWidget(Form)
self.verticalLayoutWidget.setGeometry(QtCore.QRect(13, 15, 611, 221))
self.verticalLayoutWidget.setObjectName("verticalLayoutWidget")
self.verticalLayout = QtWidgets.QVBoxLayout(self.verticalLayoutWidget)
self.verticalLayout.setSizeConstraint(QtWidgets.QLayout.SetDefaultConstraint)
self.verticalLayout.setContentsMargins(0, 0, 0, 0)
self.verticalLayout.setObjectName("verticalLayout")
self.label = QtWidgets.QLabel(self.verticalLayoutWidget)
font = QtGui.QFont()
font.setFamily("隶书")
font.setPointSize(24)
self.label.setFont(font)
self.label.setLayoutDirection(QtCore.Qt.LeftToRight)
self.label.setAlignment(QtCore.Qt.AlignCenter)
self.label.setObjectName("label")
self.verticalLayout.addWidget(self.label)
self.horizontalLayout = QtWidgets.QHBoxLayout()
self.horizontalLayout.setObjectName("horizontalLayout")
self.Show = QtWidgets.QPlainTextEdit(self.verticalLayoutWidget)
self.Show.setObjectName("Show")
self.horizontalLayout.addWidget(self.Show)
self.formLayout = QtWidgets.QFormLayout()
self.formLayout.setObjectName("formLayout")
self.class_Button = QtWidgets.QPushButton(self.verticalLayoutWidget)
self.class_Button.setMinimumSize(QtCore.QSize(50, 0))
self.class_Button.setObjectName("class_Button")
self.formLayout.setWidget(1, QtWidgets.QFormLayout.LabelRole, self.class_Button)
self.class_lineEdit = QtWidgets.QLineEdit(self.verticalLayoutWidget)
self.class_lineEdit.setObjectName("class_lineEdit")
self.formLayout.setWidget(1, QtWidgets.QFormLayout.FieldRole, self.class_lineEdit)
self.work_Button = QtWidgets.QPushButton(self.verticalLayoutWidget)
self.work_Button.setObjectName("work_Button")
self.formLayout.setWidget(2, QtWidgets.QFormLayout.LabelRole, self.work_Button)
self.work_lineEdit = QtWidgets.QLineEdit(self.verticalLayoutWidget)
self.work_lineEdit.setObjectName("work_lineEdit")
self.formLayout.setWidget(2, QtWidgets.QFormLayout.FieldRole, self.work_lineEdit)
self.run_Button = QtWidgets.QPushButton(self.verticalLayoutWidget)
self.run_Button.setObjectName("run_Button")
self.formLayout.setWidget(5, QtWidgets.QFormLayout.LabelRole, self.run_Button)
self.horizontalLayout_2 = QtWidgets.QHBoxLayout()
self.horizontalLayout_2.setObjectName("horizontalLayout_2")
self.label_2 = QtWidgets.QLabel(self.verticalLayoutWidget)
self.label_2.setObjectName("label_2")
self.horizontalLayout_2.addWidget(self.label_2)
self.class_row = QtWidgets.QLineEdit(self.verticalLayoutWidget)
self.class_row.setText("")
self.class_row.setObjectName("class_row")
self.horizontalLayout_2.addWidget(self.class_row)
self.label_3 = QtWidgets.QLabel(self.verticalLayoutWidget)
self.label_3.setObjectName("label_3")
self.horizontalLayout_2.addWidget(self.label_3)
self.woek_row = QtWidgets.QLineEdit(self.verticalLayoutWidget)
self.woek_row.setText("")
self.woek_row.setObjectName("woek_row")
self.horizontalLayout_2.addWidget(self.woek_row)
self.formLayout.setLayout(5, QtWidgets.QFormLayout.FieldRole, self.horizontalLayout_2)
self.textBrowser = QtWidgets.QTextBrowser(self.verticalLayoutWidget)
self.textBrowser.setObjectName("textBrowser")
self.formLayout.setWidget(6, QtWidgets.QFormLayout.FieldRole, self.textBrowser)
self.horizontalLayout.addLayout(self.formLayout)
self.verticalLayout.addLayout(self.horizontalLayout)
self.retranslateUi(Form)
QtCore.QMetaObject.connectSlotsByName(Form)
Form.setTabOrder(self.class_Button, self.work_Button)
def retranslateUi(self, Form):
_translate = QtCore.QCoreApplication.translate
Form.setWindowTitle(_translate("Form", "Form"))
self.label.setText(_translate("Form", "出勤名单核对系统"))
self.class_Button.setText(_translate("Form", "选择名单"))
self.work_Button.setText(_translate("Form", "选择出勤表"))
self.run_Button.setText(_translate("Form", "开始筛查"))
self.label_2.setText(_translate("Form", "名单列"))
self.class_row.setPlaceholderText(_translate("Form", "1"))
self.label_3.setText(_translate("Form", "出勤列"))
self.woek_row.setPlaceholderText(_translate("Form", "1"))
self.textBrowser.setHtml(_translate("Form", "\n"
"\n"
"使用说明：确定名册姓名所在列数，出勤姓名所在列数，填写以上位置，选择对应文件即可开始进行出勤核对
"))
能力有限仅供参考,欢迎提出宝贵意见！
