python自动化之 excel转word

docx参考手册：python文档 — python-docx 0.8.11 文档 (osgeo.cn) https://www.osgeo.cn/openpyxl/index.html

OpenPYXL参考手册：开发 — openpyxl 3.0.5 文档 (osgeo.cn) https://www.osgeo.cn/openpyxl/development.htm

需求：将一个Excel工作表的B列前5个单元格内容转为word的5个段落。

代码：

#-*- coding: utf-8 -*-
import os,openpyxl,glob
from openpyxl.styles import Border, Side,PatternFill
from docx import Document
from docx.shared import Pt,Cm,Inches
from docx.oxml.ns import qn
from docx.enum.style  import WD_STYLE_TYPE
from docx.enum.text import WD_LINE_SPACING

if __name__ == '__main__':
    #word
    Doc = Document()
    filelist=[]
    i=0
    for filename in glob.glob(r'*.xlsx'):
        filelist.append(filename)
        print(i,filename)
        i=i+1        
    f_num = input("请输入EXCEL文件序号:")
    file = filelist[int(f_num)]
    #读取EXCEL的B列前5个单元格
    wb = openpyxl.load_workbook(file)
    ws = wb.active
    i=1
    for cell in ws["B"]:
        if i>5:
            break
        i=i+1
         #存入word
        Doc.add_paragraph(str(cell.value))    
    Doc.save("example.doc")

类似的：

#-*- coding: utf-8 -*-
from openpyxl import Workbook
from openpyxl import load_workbook
import openpyxl
from docx import Document
from docx.shared import Pt,Cm
from docx.oxml.ns import qn

#global
items = []
file1 = r"F:统计表.xlsx"
#新建word
Doc = Document()
print('新建成功')

#读取数据
wb1=openpyxl.load_workbook(file1)
#获取sheet页
sheet1=wb1._sheets[0]
for m in range(1,100):
    v = sheet1["H"+str(m)].value
    if v == None:       
        pass
    else:
        w =v.split("
")
        for x in w:
            items.append(x)
wb1.close()#关闭excel
for n in range(1,len(items)):    
    Doc.add_paragraph(items[n])
#段落格式
for p in Doc.paragraphs:
    paragraph_format=p.paragraph_format
    paragraph_format.space_before=Pt(0)    #上行间距
    paragraph_format.space_after=Pt(0)    #下行间距
    paragraph_format.line_spacing=Pt(13)  #行距
    #字体
    for run in p.runs:
        run.font.size=Pt(12)
        run.font.name='宋体'
        # 设置像黑体这样的中文字体，必须添加下面 2 行代码
        r = run._element
        r.rPr.rFonts.set(qn("w:eastAsia"),"黑体")
#页边距
#changing the page margins
margin =1.0
sections = Doc.sections
for section in sections:
    section.top_margin = Cm(margin)
    section.bottom_margin = Cm(margin)
    section.left_margin = Cm(margin)
    section.right_margin = Cm(margin)    
Doc.save('test.doc')#保存数据

改成更通用的

用鼠标拖放xlsx文件，在同目录下生成docx,并限定要转换的列和行

#-*- coding: utf-8 -*-
from openpyxl import Workbook
from openpyxl import load_workbook
import openpyxl
from docx import Document
from docx.shared import Pt,RGBColor,Cm
from docx.oxml.ns import qn
from docx.enum.text import WD_ALIGN_PARAGRAPH  #设置对其方式
from tkinter import *
import tkinter
from tkinter.messagebox import showinfo
import windnd
import os

#要获取的列
cols =['O','P','I']
#要获取的最大行号
maxrow =100

#xlsx转docx
def convxlsx(file1):
    items = []    
    #新建word
    Doc = Document()
    #页边距 
    section = Doc.sections[0]    
    section.top_margin = Cm(2.54)
    section.bottom_margin = Cm(2.54)
    section.left_margin = Cm(3.18)
    section.right_margin = Cm(3.18)
    #读取数据
    wb1=openpyxl.load_workbook(file1)
    #获取sheet页
    sheet1=wb1._sheets[0]
    for m in range(1,maxrow):
        for N in cols:            
            v = sheet1[N+str(m)].value
            if v == None:       
                pass
            else:
                p = Doc.add_paragraph(v)
                p.style.font.size=Pt(16)
                p.alignment = WD_ALIGN_PARAGRAPH.LEFT
                p.paragraph_format.first_line_indent = p.style.font.size * 2
                if N=="O":
                    p.paragraph_format.space_before=Pt(0)    #上行间距
                    p.paragraph_format.space_after=Pt(0)    #下行间距
                    for run in p.runs:                        
                        run.font.name='方正楷体简体'
                        run._element.rPr.rFonts.set(qn('w:eastAsia'), '方正楷体简体')
                        run.font.color.rgb=RGBColor(0,0,0)
                        run.font.size=Pt(16)
                else:
                    p.paragraph_format.space_before=Pt(0)    #上行间距
                    p.paragraph_format.space_after=Pt(0)    #下行间距
                    for run in p.runs:                    
                        run.font.name='方正仿宋简体'
                        run._element.rPr.rFonts.set(qn('w:eastAsia'), '方正仿宋简体')
                        run.font.color.rgb=RGBColor(0,0,0)
                        run.font.size=Pt(16)                   
    wb1.close()#关闭excel
    
    file_dir =os.path.split(file1)[0]
    oldfname = os.path.split(file1)[1][:-5]
    newfname =file_dir +"\" + oldfname + '.py.docx'
    Doc.save(newfname)#保存数据
    showinfo('提示','新建成功')
              
#检查文件为xlsx
def checkxlsx(file):
    f = file[-5:]
    if f != '.xlsx':
        showinfo('错误','文件扩展名必须是xlsx')
    else:        
        convxlsx(file)

#获取拖放的文件名
def dragged_files(files):
    msg = '
'.join((item.decode('gbk') for item in files))
    #showinfo('您拖放的文件',msg)    
    file= files[0].decode('gbk')
    checkxlsx(file)
    
#创建root窗口
root = Tk()# 初始化
root.title("xlsx转doc,请将xlsx文件拖放到窗体，将在同目录下生成docx文档")
root.geometry('600x200+100+100') #长x宽+左上角x和y
windnd.hook_dropfiles(root,func=dragged_files)
root.mainloop()