递归查询树tree 根据父节点查询子集

表结构

SELECT
	ID.LEVEL1,
	DATA1.* 
FROM
	(
SELECT
	@ids AS _ids,
	( SELECT @ids := GROUP_CONCAT( pid ) FROM project_grouping WHERE FIND_IN_SET( parent_id, @ids ) ) AS cids,
	@l := @l + 1 AS LEVEL1 
FROM
	project_grouping,
	( SELECT @ids := 1, @l := 0 ) b 
WHERE
	@ids IS NOT NULL 
	) ID,
	project_grouping DATA1 
WHERE
	FIND_IN_SET( DATA1.pid, ID._ids ) 
ORDER BY
	LEVEL1,
	pid

不太理解网上为什么用LEVEL当作是别名 明明是关键字 

@ids 为父节点id

由于业务中经常使用tree来返回而不是list<对象的形式>，现在java中进行递归造json tree

@Override
	public Object selectGrouping() {
		 List<Project_grouping> list=Project_groupingMapper.selectTreeByParent_id(0,null);
		return	buildTree(list,0);
	}

 public static List<TreeGroup> buildTree(List<Project_grouping> list,int parent_id){
		List<TreeGroup> trees =new ArrayList<TreeGroup>();
		   for (Project_grouping group : list) {
		          int id = group.getGid();//主键id
		          int pid = group.getParent_id();//父类id
		          String name = group.getName();//名称
		          if (parent_id == pid) {  
		              List<TreeGroup> treeLists = buildTree(list, id);  
		                TreeGroup tree = new TreeGroup(id, parent_id, name, treeLists); 
		                trees.add(tree);  
		          }  
		      }  
		     
		return trees;
	}

        private Integer id;
	 
	    private Integer parent_id;
	 
	    private String name;
	    
	    private List<TreeGroup> children;
	    
	    public TreeGroup(Integer id, Integer parent_id, String name, List<TreeGroup> children) {
			super();
			this.id = id;
			this.parent_id = parent_id;
			this.name = name;
			this.children = children;
		}