最近开发过程中遇到一个Excel的导入的功能，因为导入的数据结构具有层次结构，经过一番研究，最终得以实现，所有写下该文章，记录过程，供以后参考。
下图是导入Excel的数据结构：
使用POI解析Excel，数据封装然后进行入库。下面是核心代码。
@Override
public KnowledgeBaseDictImportRespVO importKnowledgeBaseDict(MultipartFile file) throws IOException {
KnowledgeBaseDictImportRespVO kbdir = KnowledgeBaseDictImportRespVO.builder().build(); // 该对象是方法返回的结果对象
Map<String,String> res = new ConcurrentHashMap<>();
List<String> createProblemsList = new ArrayList<>();
String fileName = file.getOriginalFilename();
String fileType = fileName.substring(fileName.lastIndexOf(".")).toLowerCase();
if(fileType.equals(".xls") || fileType.equals(".xlsx")) {
AdminUserRespDTO adminUserRespDTO = adminUserApi.getUser(SecurityFrameworkUtils.getLoginUserId()).getData();
//poi解析excel
InputStream inputStream = file.getInputStream();
// 读取工作薄 .xls 与 .xlsx 需要创建不通的工作簿 采用工作簿工厂类创建对应的工作簿类
Workbook workbook = WorkbookFactory.create(inputStream);
Sheet sheet = workbook.getSheetAt(0);
//获取系统名称
Row row0 = sheet.getRow(0); // 第一行的表头
Cell cell0 = row0.getCell(0); // 第一列
String belongSystem = cell0.getStringCellValue().trim().replace("系统功能树","");
if (StringUtils.isEmpty(belongSystem)) {
res.put(StrUtils.getRandomString(6),"归属系统不能为空!");
kbdir.setFailureProblems(res);
return kbdir;
}
KnowledgeBaseDictDO kbdictdo = knowledgeBaseDictMapper.getKnowledgeBaseDictByParenetIdAndMenuName(0L,belongSystem);
//定义一个cellId,此为每一次循环前一列的id
long cellId = 0;
String cellName = null; // 记录父级名称
String firstMenuName = null; // 因为合并单元格只有第一个有值,需要记录该值。
int count = 0;
if (ObjectUtils.isEmpty(kbdictdo)) {
kbdictdo = new KnowledgeBaseDictDO();
kbdictdo.setSort(0);
kbdictdo.setParentId(0L);
kbdictdo.setMenuLevel(0);
kbdictdo.setMenuName(belongSystem);
kbdictdo.setBelongingSystem(belongSystem);
kbdictdo.setCreator(adminUserRespDTO.getNickname());
int rs = knowledgeBaseDictMapper.insert(kbdictdo);
if (rs > 0) {
count += rs;
createProblemsList.add(belongSystem);
cellId = kbdictdo.getId();
cellName = kbdictdo.getMenuName();
}
}
int rows = sheet.getPhysicalNumberOfRows();
if (rows > 2) {
//按照行进行循环，读取当前行的列
for (int i = 0,j = 2; j < rows; j++,i++) {
// 读取行
Row row = sheet.getRow(j);
//查询当前行有多少列
int physical = sheet.getRow(j).getPhysicalNumberOfCells();
if (row != null) {
//获取第一列的单元格
Cell cell = row.getCell(0);
//判断单元格是否为空
if (!(cell == null || "".equals(cell.toString().trim()))) {
firstMenuName = cell.getStringCellValue().trim();
//根据单元格的数据查询数据库是否存在记录
KnowledgeBaseDictDO kbddoFirst = knowledgeBaseDictMapper.getKnowledgeBaseDictByMenuNameAndMenuLevelAndbls(firstMenuName,1,belongSystem);
//数据库不存在，添加记录
if (ObjectUtils.isEmpty(kbddoFirst)) {
kbddoFirst = new KnowledgeBaseDictDO();
kbddoFirst.setSort(i);
kbddoFirst.setParentId(cellId);
kbddoFirst.setBelongingSystem(belongSystem);
kbddoFirst.setParentMenuName(cellName);
kbddoFirst.setMenuName(cell.getStringCellValue().trim());
kbddoFirst.setMenuLevel(1);
kbddoFirst.setCreator(adminUserRespDTO.getNickname());
int rs = knowledgeBaseDictMapper.insert(kbddoFirst);
if (rs > 0) {
count += rs;
createProblemsList.add(cell.getStringCellValue().trim());
// 记录父级id 与 名称
cellId = kbddoFirst.getId();
cellName = kbddoFirst.getMenuName();
}
} else {
//数据库存在记录，将这条记录的id作为父id
cellId = kbddoFirst.getId();
cellName = kbddoFirst.getMenuName();
}
} else { // 表示循环到合并单元格的地方,查询数据库记录父级信息
if (StringUtils.isNotEmpty(firstMenuName)) {
KnowledgeBaseDictDO kbddoFirst = knowledgeBaseDictMapper.getKnowledgeBaseDictByMenuNameAndMenuLevelAndbls(firstMenuName,1,belongSystem);
cellId = kbddoFirst.getId();
cellName = kbddoFirst.getMenuName();
}
}
}
// 从第二列单元格开始
for (int k = 1,l = 0; k < physical; k++,l++) {
//取单元格
Cell cell = row.getCell(k);
//判断单元格是否为空
if (!(cell == null || "".equals(cell.toString().trim()))) {
//查询数据库有无此记录
KnowledgeBaseDictDO kbddoSecond = knowledgeBaseDictMapper.getKnowledgeBaseDictByMenuNameAndMenuLevelAndbls(cell.getStringCellValue().trim(),k + 1,belongSystem);
if (ObjectUtils.isEmpty(kbddoSecond)) {
kbddoSecond = new KnowledgeBaseDictDO();
kbddoSecond.setSort(l);
kbddoSecond.setParentId(cellId);
kbddoSecond.setBelongingSystem(belongSystem);
kbddoSecond.setParentMenuName(cellName);
kbddoSecond.setMenuName(cell.getStringCellValue().trim());
kbddoSecond.setMenuLevel(k+1);
kbddoSecond.setCreator(adminUserRespDTO.getNickname());
int rs = knowledgeBaseDictMapper.insert(kbddoSecond);
if (rs > 0) {
count += rs;
createProblemsList.add(cell.getStringCellValue().trim());
// 记录父级id 与 名称
cellId = kbddoSecond.getId();
cellName = kbddoSecond.getMenuName();
}
} else {
cellId = kbddoSecond.getId();
cellName = kbddoSecond.getMenuName();
}
if (k == physical - 1) {
// 内层循环结束,表示一行的一级菜单结束,需要指定一级菜单的父级
cellId = kbdictdo.getId();
cellName = kbdictdo.getMenuName();
}
} else { // 如果为空,需要记录父级id与名称
cellId = kbdictdo.getId();
cellName = kbdictdo.getMenuName();
}
}
}
} else {
res.put(StrUtils.getRandomString(6),"导入的数据不能为空!");
kbdir.setFailureProblems(res);
return kbdir;
}
kbdir.setCreateProblems(createProblemsList);
kbdir.setFailureProblems(res);
}
return kbdir;
}该方法也适用于下面的Excel数据结构导入（未进行列合并单元格的数据结构）：
树形结构数据如何高效封装，通过接口返回前端展示，请看我的另一篇文章。
Java 树形结构数据如何高效返回给前端进行展示？-CSDN博客
