'commit'

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

@@ -28,43 +28,6 @@ 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";
 
-    /**
-     * 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);
-        is.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
-        int totalRow = data.size() + skipRowCount;
-        if (totalRow < expectLimit) {
-            System.out.println("期望数量>=" + expectLimit + ",现在只有" + totalRow + "条，理解为POI读取WORD图表存在问题，使用python进行二次获取数据...");
-            // 留出足够的com关闭word的时间长度，否则会有异常
-            Thread.sleep(3000);
-            //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);
-            System.out.println("二次获取数据条目数量：" + data.size() + ",期望数量=" + expectLimit);
-        }
-        return data;
-    }
 
     public static void main(String[] args) throws IOException, InvalidFormatException, InterruptedException {
         //初始化数据库连接
@@ -104,14 +67,17 @@ public class C9 {
                     System.out.println("正在处理" + cityName + "市州文件...");
                     //数据在图表1
                     int chartNumber = 1;
-                    List<List<String>> source = getChartData(file.getAbsolutePath(), chartNumber, 6, 20);//2017年开始
+                    //chartNumber:第几个图表
+                    //skipRowCount:跳过的行数,2017年开始,第一行是表头，第二行开始是2022，所以填写了跳过6行
+                    //expectLimit:期望的数据行数,首先用POI进行解析，如果获取的行数大于预期行数，就是正确的，否则就需要二次调用python进行读取
+                    List<List<String>> source = ExcelKit.getChartData(file.getAbsolutePath(), chartNumber, 6, 20);
 
                     //遍历source1
                     for (List<String> r : source) {
                         Row outRow = outSheet.createRow(++rowIndex);
                         // 导出数据
                         //上级行政区划，行政区划，年份，总人口变化，总人口预测
-                        int year = Integer.parseInt(r.getFirst());
+                        int year = Integer.parseInt(r.getFirst().substring(0, 4));
 
                         double value = Double.parseDouble(r.get(1));
                         if (year <= 2023) {

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

@@ -2,16 +2,20 @@ package com.dsideal.base.Tools.FillData.ExcelKit;
 
 import com.dsideal.base.DataEase.Model.ExcelReader;
 import com.jfinal.kit.StrKit;
+import org.apache.poi.openxml4j.exceptions.InvalidFormatException;
+import org.apache.poi.openxml4j.util.ZipSecureFile;
 import org.apache.poi.ss.usermodel.*;
 import org.apache.poi.xssf.usermodel.XSSFCellStyle;
 import org.apache.poi.xssf.usermodel.XSSFFont;
 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.*;
 import java.nio.charset.StandardCharsets;
 import java.util.ArrayList;
+import java.util.Iterator;
 import java.util.List;
 
 public class ExcelKit {
@@ -380,4 +384,104 @@ public class ExcelKit {
         writer.flush();
         writer.close();
     }
+
+    /**
+     * 读取excel文件
+     * @param excelFilePath
+     * @param skipRows
+     * @return
+     */
+    public static List<List<String>> readExcelToList(String excelFilePath, int skipRows) {
+        List<List<String>> data = new ArrayList<>();
+        try (FileInputStream inputStream = new FileInputStream(excelFilePath)) {
+            Workbook workbook = new XSSFWorkbook(inputStream);
+            Sheet sheet = workbook.getSheetAt(0); // 获取第一个Sheet
+
+            // 创建一个迭代器来遍历行，跳过指定数量的行
+            Iterator<Row> rowIterator = sheet.iterator();
+            while (skipRows > 0 && rowIterator.hasNext()) {
+                rowIterator.next();
+                skipRows--;
+            }
+
+            // 遍历每一行
+            while (rowIterator.hasNext()) {
+                Row row = rowIterator.next();
+                List<String> rowData = new ArrayList<>();
+                // 遍历每一行中的每一列
+                Iterator<Cell> cellIterator = row.cellIterator();
+
+                while (cellIterator.hasNext()) {
+                    Cell cell = cellIterator.next();
+                    // 根据单元格的不同类型获取数据
+                    switch (cell.getCellType()) {
+                        case STRING:
+                            rowData.add(cell.getStringCellValue());
+                            break;
+                        case NUMERIC:
+                            if (DateUtil.isCellDateFormatted(cell)) {
+                                rowData.add(cell.getDateCellValue().toString());
+                            } else {
+                                rowData.add(Double.toString(cell.getNumericCellValue()));
+                            }
+                            break;
+                        case BOOLEAN:
+                            rowData.add(Boolean.toString(cell.getBooleanCellValue()));
+                            break;
+                        case FORMULA:
+                            rowData.add(cell.getCellFormula());
+                            break;
+                        default:
+                            rowData.add("");
+                            break;
+                    }
+                }
+                data.add(rowData);
+            }
+            workbook.close();
+        } catch (Exception e) {
+            e.printStackTrace();
+        }
+        return data;
+    }
+
+    /**
+     * 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);
+        is.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
+        int totalRow = data.size() + skipRowCount;
+        if (totalRow < expectLimit) {
+            System.out.println("期望数量>=" + expectLimit + ",现在只有" + totalRow + "条，理解为POI读取WORD图表存在问题，使用python进行二次获取数据...");
+            // 留出足够的com关闭word的时间长度，否则会有异常
+            Thread.sleep(4000);
+            //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
+            //使用POI
+            int skipRows = 1; // 假设我们要跳过第一行
+            data = ExcelKit.readExcelToList(ExcelKit.excelPath, skipRows);
+            System.out.println("二次获取数据条目数量：" + data.size() + ",期望数量=" + expectLimit);
+        }
+        return data;
+    }
 }