在处理组织架构树形展示时，原始代码使用了两次SQL查询，并通过Java代码进行数据组装。这种做法不仅增加了数据库访问次数，也提高了程序复杂度。

以下是优化前的实现方式：

public List<TdDepartment> createZtreeDep(String compId) {
    List<TdDepartment> result = new ArrayList<TdDepartment>();
    
    String childSql = "select dep_id,dep_name,super_id,folder from td_department " +
                      "start with super_id in ( " +                        
                      "select dep_id  from td_department " + 
                      "where valid_flag = 'Y'  and comp_id = '"+compId+"') connect by prior dep_id = super_id";
    
    String parentSql = "select dep_id ,dep_name,super_id,folder  from td_department " +
                       "where valid_flag = 'Y'  and comp_id = '"+compId+"'";
    
    EpDB db = new EpDB();        
    ArrayList<HashMap> parentDepts = db.getHashData(parentSql);
    ArrayList<HashMap> childDepts = db.getHashData(childSql);
    
    if(parentDepts == null || parentDepts.size() <= 0)
        return null;
    
    for(int i=0; i<parentDepts.size(); i++){
        String id = parentDepts.get(i).get("DEP_ID").toString();
        String name = parentDepts.get(i).get("DEP_NAME").toString();
        String pid = parentDepts.get(i).get("SUPER_ID").toString();            
        TdDepartment dept = new TdDepartment();
        dept.setId(id);
        dept.setPid(pid);
        dept.setName(name);
        if(parentDepts.get(i).get("FOLDER") != null){
            String folder = parentDepts.get(i).get("FOLDER").toString();
            if("Y".equals(folder)){
                dept.setOpen("true");
            }else{
                dept.setOpen("false");
            }
        }
        result.add(dept);
    }
    
    for(int i=0; i<childDepts.size(); i++){
        String id = childDepts.get(i).get("DEP_ID").toString();
        String name = childDepts.get(i).get("DEP_NAME").toString();
        String pid = childDepts.get(i).get("SUPER_ID").toString();            
        TdDepartment dept = new TdDepartment();
        dept.setId(id);
        dept.setPid(pid);
        dept.setName(name);
        if(childDepts.get(i).get("FOLDER") != null){
            String folder = childDepts.get(i).get("FOLDER").toString();
            if("Y".equals(folder)){
                dept.setOpen("true");
            }else{
                dept.setOpen("false");
            }
        }
        result.add(dept);
    }
    
    return result;
}

经过分析发现，可以合并两个查询为一个层级查询语句，从而减少数据库访问次数并简化逻辑。

优化后的代码如下：

public List<TdDepartment> createZtreeDep(String compId) {
    List<TdDepartment> result = new ArrayList<TdDepartment>();
    
    String sql = "select dep_id,dep_name,super_id,folder from td_department " +
                "start with dep_id in ( " +                        
                "select dep_id  from td_department " + 
                "where valid_flag = 'Y'  and comp_id = '"+compId+"') connect by super_id = prior dep_id";
    
    System.out.println("sql=" + sql);
    EpDB db = new EpDB();        
    ArrayList<HashMap> departments = db.getHashData(sql);
    
    if(departments == null || departments.size() <= 0)
        return null;
    
    for(int i=0; i<departments.size(); i++){
        String id = departments.get(i).get("DEP_ID").toString();
        String name = departments.get(i).get("DEP_NAME").toString();
        String pid = departments.get(i).get("SUPER_ID").toString();            
        TdDepartment dept = new TdDepartment();
        dept.setId(id);
        dept.setPid(pid);
        dept.setName(name);
        if(departments.get(i).get("FOLDER") != null){
            String folder = departments.get(i).get("FOLDER").toString();
            if("Y".equals(folder)){
                dept.setOpen("true");
            }else{
                dept.setOpen("false");
            }
        }
        result.add(dept);
    }
    
    return result;
}

主要改进点包括：

• 将原本两次数据库查询合并为一次
• 简化了Java代码逻辑结构
• 避免重复遍历和对象创建