Merge branch 'main' of http://10.10.14.176:3000/huanghai/YunNanDsBase

Py/RepairTuBiao.py

@@ -0,0 +1,51 @@
+# pip install pywin32
+# https://blog.csdn.net/weixin_42927998/article/details/115086797
+import os
+
+import win32com
+from win32com.client import Dispatch
+
+# 工作目录
+workingPath = r'D:\dsWork\YunNanDsBase\Doc\全省及州市县区人口与教育报告集20241023\16个州市报告2022\分析报告20240510'
+
+
+# 修复Word文档
+# 经过反复测试发现，WORD文档中的图表，有些POI是无法正确读取的，本来是Sheet1,结果它不认识，说只有一个Sheet0,此时就无法正确读取数据了。
+# 而我通过python+win32com.client.Dispatch可以读取到，直接保存，就修复了这个BUG，真是太神奇了！
+def repairWord(docPath):
+    docApp = win32com.client.Dispatch('Word.Application')
+    # 是不是打Word显示
+    docApp.Visible = False
+    docApp.DisplayAlerts = 0
+
+    doc = docApp.Documents.Open(docPath)
+    #
+    # # 遍历文档中的所有内嵌形状
+    idx = 1
+    for inline_shape in doc.InlineShapes:
+        if inline_shape.Type == win32com.client.constants.wdInlineShapeChart:  # 检查是否为内嵌图表
+            shape = doc.InlineShapes(idx)
+            sheet = shape.Chart.ChartData.Workbook.Worksheets("Sheet1")
+            # 下一个图表的索引号
+            idx = idx + 1
+
+    # 关闭文档和Word应用
+    doc.Close()
+    docApp.Quit()
+
+
+if __name__ == '__main__':
+    # 1、修复两层扩展名.docx
+    for file in os.listdir(workingPath):
+        if file.endswith('.docx.docx'):
+            # 完整的路径名称
+            docPath = os.path.join(workingPath, file)
+            print("文件名有误，已修复：" + docPath)
+            os.rename(docPath, docPath.replace('.docx.docx', '.docx'))
+
+    # 2、修复图表异常问题
+    for file in os.listdir(workingPath):
+        if file.endswith('.docx'):
+            # 开始修复文档
+            repairWord(docPath)
+    print("修复完成")

Py/Test/DocxTuBiaoRead.py

@@ -7,6 +7,20 @@ docApp = win32com.client.Dispatch('Word.Application')
 # 是不是打Word显示
 docApp.Visible = False
 docApp.DisplayAlerts = 0
+<<<<<<< HEAD
+
+doc = docApp.Documents.Open("c:/b.docx")
+#
+# # 遍历文档中的所有内嵌形状
+idx = 1
+for inline_shape in doc.InlineShapes:
+    if inline_shape.Type == win32com.client.constants.wdInlineShapeChart:  # 检查是否为内嵌图表
+        # 获取图表的标题,此项目中图表没有标题
+        shape = doc.InlineShapes(idx)
+        sheet = shape.Chart.ChartData.Workbook.Worksheets("Sheet1")
+        # 下一个图表的索引号
+        idx = idx + 1
+=======
 working_dir = r"D:/dsWork/YunNanDsBase/Doc/全省及州市县区人口与教育报告集20241023/16个州市报告2022/分析报告20240510/"
 
 # doc = docApp.Documents.Open('c:/1.docx')
@@ -45,6 +59,7 @@ for inline_shape in doc.InlineShapes:
         idx = idx + 1
 print(idx-1)
 
+>>>>>>> cfa48d4d9b83ad3f82bc55fd60f8926d113dfe09
 
 # 关闭文档和Word应用
 doc.Close()

src/main/java/com/dsideal/base/Test/Test1.java

@@ -0,0 +1,56 @@
+package com.dsideal.base.Test;
+
+import org.apache.poi.xssf.usermodel.XSSFRow;
+import org.apache.poi.xssf.usermodel.XSSFSheet;
+import org.apache.poi.xssf.usermodel.XSSFWorkbook;
+import org.apache.poi.xwpf.usermodel.XWPFDocument;
+
+
+import org.apache.poi.xwpf.usermodel.XWPFChart;
+import org.openxmlformats.schemas.drawingml.x2006.chart.CTChart;
+
+import org.openxmlformats.schemas.drawingml.x2006.chart.CTSerTx;
+import org.openxmlformats.schemas.drawingml.x2006.chart.CTTitle;
+import org.openxmlformats.schemas.spreadsheetml.x2006.main.CTWorkbook;
+
+import java.io.FileInputStream;
+import java.util.List;
+
+public class Test1 {
+    public static void main(String[] args) throws Exception {
+        FileInputStream fis = new FileInputStream("c:\\b.docx");
+        XWPFDocument document = new XWPFDocument(fis);
+
+        // 获取文档中的第一个图表
+        List<XWPFChart> shapes = document.getCharts();
+
+
+
+        for (int i = 0; i < shapes.size(); i++) {
+
+            XSSFWorkbook workbook = shapes.get(i).getWorkbook();
+            XSSFSheet sheet = workbook.getSheet(workbook.getSheetName(0));
+            if(i==21)
+            {
+                System.out.println(11111);
+            }
+
+            XSSFRow row = sheet.getRow(0);
+            System.out.println(row);
+            System.out.println(row.getCell(0));
+            System.out.println(i);
+        }
+
+
+
+
+
+
+
+
+
+
+
+        fis.close();
+    }
+}