修复在mac 2016 2017导出的excel 可能存在多余字段的问题

alibaba · Sep 5, 2019 · cd2e0fe · cd2e0fe
1 parent 120dfbd
commit cd2e0fe
Show file tree

Hide file tree

Showing 4 changed files with 26 additions and 3 deletions.
diff --git a/src/main/java/com/alibaba/excel/analysis/v07/SharedStringsTableHandler.java b/src/main/java/com/alibaba/excel/analysis/v07/SharedStringsTableHandler.java
@@ -13,6 +13,11 @@
 public class SharedStringsTableHandler extends DefaultHandler {
     private static final String T_TAG = "t";
     private static final String SI_TAG = "si";
+    /**
+     * Mac 2016 2017 will have this extra field to ignore
+     */
+    private static final String RPH_TAG = "rPh";
+
     /**
      * The final piece of data
      */
@@ -23,6 +28,14 @@ public class SharedStringsTableHandler extends DefaultHandler {
     private StringBuilder currentElementData;
 
     private ReadCache readCache;
+    /**
+     * Some fields in the T tag need to be ignored
+     */
+    private boolean ignoreTagt = false;
+    /**
+     * The only time you need to read the characters in the T tag is when it is used
+     */
+    private boolean isTagt = false;
 
     public SharedStringsTableHandler(ReadCache readCache) {
         this.readCache = readCache;
@@ -32,8 +45,11 @@ public SharedStringsTableHandler(ReadCache readCache) {
     public void startElement(String uri, String localName, String name, Attributes attributes) {
         if (T_TAG.equals(name)) {
             currentElementData = null;
+            isTagt = true;
         } else if (SI_TAG.equals(name)) {
             currentData = null;
+        } else if (RPH_TAG.equals(name)) {
+            ignoreTagt = true;
         }
     }
 
@@ -46,17 +62,23 @@ public void endElement(String uri, String localName, String name) {
                 }
                 currentData.append(currentElementData);
             }
+            isTagt = false;
         } else if (SI_TAG.equals(name)) {
             if (currentData == null) {
                 readCache.put(null);
             } else {
                 readCache.put(currentData.toString());
             }
+        } else if (RPH_TAG.equals(name)) {
+            ignoreTagt = false;
         }
     }
 
     @Override
     public void characters(char[] ch, int start, int length) {
+        if (!isTagt || ignoreTagt) {
+            return;
+        }
         if (currentElementData == null) {
             currentElementData = new StringBuilder();
         }

diff --git a/src/test/java/com/alibaba/easyexcel/test/temp/poi/PoiFormatTest.java b/src/test/java/com/alibaba/easyexcel/test/temp/poi/PoiFormatTest.java
@@ -28,7 +28,7 @@ public class PoiFormatTest {
 
     @Test
     public void lastRowNum() throws IOException {
-        String file = TestFileUtil.getPath() + "demo" + File.separator + "demo.xlsx";
+        String file = "D:\\test\\原文件.xlsx";
         SXSSFWorkbook xssfWorkbook = new SXSSFWorkbook(new XSSFWorkbook(file));
         SXSSFSheet xssfSheet = xssfWorkbook.getSheetAt(0);
         LOGGER.info("一共行数:{}", xssfSheet.getLastRowNum());

diff --git a/src/test/java/com/alibaba/easyexcel/test/temp/simple/HgTest.java b/src/test/java/com/alibaba/easyexcel/test/temp/simple/HgTest.java
@@ -23,8 +23,8 @@ public class HgTest {
 
     @Test
     public void hh() throws IOException {
-        List<Object> list = EasyExcel.read(new FileInputStream("D:\\test\\hg2.xls")).autoTrim(Boolean.FALSE).sheet(2)
-            .headRowNumber(0).doReadSync();
+        List<Object> list =
+            EasyExcel.read(new FileInputStream("D:\\test\\原文件.xlsx")).headRowNumber(0).sheet().doReadSync();
         for (Object data : list) {
             LOGGER.info("返回数据：{}", JSON.toJSONString(data));
         }

diff --git a/update.md b/update.md
@@ -1,5 +1,6 @@
 # 2.0.0-beta4
 * 修改在传入List<List<Object>>判断行数错误 [Issue #526](https://github.com/alibaba/easyexcel/issues/526)
+* 修复在mac 2016 2017导出的excel 可能存在多余字段的问题
 
 # 2.0.0-beta3
 * 导出完成移除临时目录 [Issue #386](https://github.com/alibaba/easyexcel/issues/386)