'commit'

Py/TuBiao.py

@@ -1,11 +1,9 @@
 # pip install pywin32
 # https://blog.csdn.net/weixin_42927998/article/details/115086797
-import os
 
 import win32com
-from win32com.client import constants, Dispatch
-
 from openpyxl import Workbook
+from win32com.client import constants
 
 if __name__ == '__main__':
     # 文件路径

src/main/java/com/dsideal/base/Tools/FillData/City/C9.java

@@ -28,7 +28,41 @@ public class C9 {
     //示例Excel
     static String sampleExcelPath = "D:\\dsWork\\YunNanDsBase\\Doc\\待处理\\市\\【9】总人口变化及预测-双\\总人口变化及预测-双.xlsx";
 
-    public static void main(String[] args) throws IOException, InvalidFormatException {
+    /**
+     * 获取指定文档中指定图表数据
+     *
+     * @param docPath      文档路径
+     * @param chartNumber  图表序号
+     * @param skipRowCount 跳过的行数
+     * @return 结果数据
+     * @throws IOException
+     * @throws InvalidFormatException
+     */
+    public static List<List<String>> getChartData(String docPath, int chartNumber, int skipRowCount, int expectLimit) throws IOException, InvalidFormatException, InterruptedException {
+        InputStream is = new FileInputStream(docPath);
+        ZipSecureFile.setMinInflateRatio(-1.0d);
+        XWPFDocument doc = new XWPFDocument(is);
+        //排序后的图表
+        List<XWPFChart> charts = ExcelKit.getSortListForXWPFChart(doc.getCharts());
+        XSSFWorkbook workbook = charts.get(chartNumber).getWorkbook();
+        List<List<String>> data = ExcelKit.readSheet(workbook, skipRowCount);
+        workbook.close();
+
+        //如果达到目标预期的数量，就直接返回poi获取的数据列表
+        if (data.size() < expectLimit) {
+            System.out.println("数据不足，重新获取数据,现正在使用python_docx进行二次获取数据...");
+            //否则调用python+com进行再次获取数据列表，这次获取的可能才是对的
+            //写入交互文本文件
+            ExcelKit.callPythonPrepare(docPath, chartNumber);
+            //对图表进行读取
+            ExcelKit.callPythonRead();
+            //读取生成的EXCEL,使用POI就可以了
+            data = ExcelKit.readSheet(ExcelKit.excelPath, skipRowCount);
+        }
+        return data;
+    }
+
+    public static void main(String[] args) throws IOException, InvalidFormatException, InterruptedException {
         //初始化数据库连接
         LocalMysqlConnectUtil.Init();
         //实例化
@@ -65,25 +99,16 @@ public class C9 {
                 if (fileName.endsWith(".docx") && !fileName.startsWith("~")) {
                     System.out.println("正在处理" + cityName + "市州文件...");
                     //读取文件
-                    String inputUrl = file.getAbsolutePath();
-                    InputStream is = new FileInputStream(inputUrl);
-                    ZipSecureFile.setMinInflateRatio(-1.0d);
-                    XWPFDocument doc = new XWPFDocument(is);
-                    //排序后的图表
-                    List<XWPFChart> charts = ExcelKit.getSortListForXWPFChart(doc.getCharts());
+
 
                     //数据在图表1
                     int firstChartNumber = 1;
+                    List<List<String>> source1 = getChartData(file.getAbsolutePath(), firstChartNumber, 6,20);//2017年开始
 
-                    XSSFWorkbook workbook = charts.get(firstChartNumber).getWorkbook();
-//                    if(cityName.contains("丽江")){
-//                        System.out.println("丽江");
-//                    }
-                    List<List<String>> source1 = ExcelKit.readSheet(workbook, 6);//从2017年开始
-
-                    System.out.println(source1);
-
-
+                    if (cityName.contains("丽江")) {
+                        System.out.println("丽江");
+                        System.out.println(source1);
+                    }
 
                     //遍历source1
                     for (List<String> r : source1) {

src/main/java/com/dsideal/base/Tools/FillData/ExcelKit/ExcelKit.java

@@ -16,6 +16,13 @@ import java.util.List;
 
 public class ExcelKit {
 
+    //与python交互使用的excel文件路径
+    public static String excelPath = "c:/task.xlsx";
+    //执行的python路径，这里我使用的是anaconda3的python，路径自行修改,注意要在这个环境中pip安装了python-docx，否则会报错
+    public static String python = "D:\\anaconda3\\envs\\py310\\python.exe";
+    //python脚本路径
+    public static String py = "D:\\dsWork\\YunNanDsBase\\Py\\TuBiao.py";
+
     /**
      * 将xls转换为xlsx
      *
@@ -143,7 +150,7 @@ public class ExcelKit {
         // 遍历工作表中的所有行
         if (sheet == null) return array;
 
-        System.out.println("Sheet rows="+sheet.getPhysicalNumberOfRows());
+        System.out.println("Sheet rows=" + sheet.getPhysicalNumberOfRows());
         for (Row row : sheet) {
             rowIndex++;
             if (rowIndex <= skipRowCount) continue;//跳过指定的行数
@@ -163,7 +170,6 @@ public class ExcelKit {
             }
             array.add(X);
         }
-        workbook.close();
         return array;
     }
 
@@ -333,6 +339,7 @@ public class ExcelKit {
         }
         return --rlt;
     }
+
     /**
      * 调用python+com读取WORD中的图表
      *
@@ -340,9 +347,9 @@ public class ExcelKit {
      * @throws InterruptedException
      */
     public static void callPythonRead() throws IOException, InterruptedException {
+
+        ExcelKit.delExcel(excelPath);
         // 创建ProcessBuilder对象，并设置Python脚本的路径
-        String python = "D:\\anaconda3\\envs\\py310\\python.exe";
-        String py = "D:\\dsWork\\YunNanDsBase\\Py\\TuBiao.py";
         ProcessBuilder processBuilder = new ProcessBuilder(python, py);
         // 重定向错误流到标准输出，这样可以在Java中捕获所有的输出
         processBuilder.redirectErrorStream(true);
@@ -358,7 +365,7 @@ public class ExcelKit {
         process.waitFor();
     }
 
-    public static void callPythonPrepare(String docPath,int tuBiaoNum) throws IOException {
+    public static void callPythonPrepare(String docPath, int tuBiaoNum) throws IOException {
         String taskTxt = "c:/task.txt";
         //如果文件存在则删除
         if (new File(taskTxt).exists()) {

src/main/java/com/dsideal/base/Tools/FillData/Test/TestBadExcel.java

@@ -0,0 +1,71 @@
+package com.dsideal.base.Tools.FillData.Test;
+
+import cn.hutool.core.io.FileUtil;
+import com.dsideal.base.Tools.FillData.ExcelKit.ExcelKit;
+import com.dsideal.base.Tools.Util.LocalMysqlConnectUtil;
+import com.dsideal.base.Tools.Util.ReadDocxUtil;
+import org.apache.poi.openxml4j.exceptions.InvalidFormatException;
+import org.apache.poi.openxml4j.util.ZipSecureFile;
+import org.apache.poi.ss.usermodel.Row;
+import org.apache.poi.xssf.usermodel.XSSFCellStyle;
+import org.apache.poi.xssf.usermodel.XSSFSheet;
+import org.apache.poi.xssf.usermodel.XSSFWorkbook;
+import org.apache.poi.xwpf.usermodel.XWPFChart;
+import org.apache.poi.xwpf.usermodel.XWPFDocument;
+
+import java.io.File;
+import java.io.FileInputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.List;
+
+public class TestBadExcel {
+    //开始读取市州word文档
+    static String parentPath = "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";
+
+    //示例Excel
+    static String sampleExcelPath = "D:\\dsWork\\YunNanDsBase\\Doc\\待处理\\市\\【11】教育资源配置发展预测\\教育资源配置发展预测(人).xlsx";
+
+    public static void main(String[] args) throws IOException, InvalidFormatException {
+        //初始化数据库连接
+        LocalMysqlConnectUtil.Init();
+        //实例化
+        ReadDocxUtil ru = new ReadDocxUtil();
+
+
+        //找到parentPath下一级目录中所有文件
+        List<File> files = FileUtil.loopFiles(parentPath, file -> true);
+        int rowIndex = 0;
+        //处理这个目录
+        if (files != null) {
+            for (File file : files) {
+                //判断file是不是目录，是目录的需要跳过
+                if (file.isDirectory()) continue;
+                //城市名称
+                String cityName = ru.getCityOrAreaName(file.getName());
+                String fileName = file.getName();
+
+                //判断是否为docx文件
+                if (fileName.endsWith(".docx") && !fileName.startsWith("~")) {
+                    System.out.println("正在处理" + cityName + "市州文件...");
+                    //读取文件
+                    String inputUrl = file.getAbsolutePath();
+                    InputStream is = new FileInputStream(inputUrl);
+                    ZipSecureFile.setMinInflateRatio(-1.0d);
+                    XWPFDocument doc = new XWPFDocument(is);
+                    //排序后的图表
+                    List<XWPFChart> charts = ExcelKit.getSortListForXWPFChart(doc.getCharts());
+
+                    //数据在图表36，教职工总量
+                    int firstChartNumber = 36;
+                    if(cityName.contains("西双版纳州")){
+                        System.out.println("he");
+                    }
+                    charts.get(firstChartNumber - 1).getWorkbook();
+                }
+            }
+        }
+    }
+}