如何在Neo4j中将带间隔的有序整数集合拆分为连续范围？

可以通过Cypher的窗口函数（LAG、SUM OVER）实现类似map/reduce的分组逻辑，将连续整数归为一组后聚合生成范围字符串，具体步骤如下：

核心逻辑 #

1. 将有序集合拆分为单个元素，逐个对比相邻元素的差值；
2. 标记差值大于1的位置为新分组起点；
3. 通过累加标记值生成分组ID，把连续元素归为同一组；
4. 按分组ID聚合，取每组的最小/最大值拼接成范围格式。

完整Cypher示例 #

// 替换为你的有序整数集合（比如聚合查询结果）
WITH [1,2,3,4,5,7,8,9,10,45,46,47,48] AS nums
UNWIND nums AS num
// 标记新分组起点：第一个元素 或 当前元素与前一个差值>1
WITH 
  num,
  CASE 
    WHEN LAG(num) OVER () IS NULL OR num - LAG(num) OVER () > 1 
    THEN 1 
    ELSE 0 
  END AS is_new_group
// 累加标记生成分组ID，连续元素将获得相同ID
WITH 
  num,
  SUM(is_new_group) OVER (ORDER BY num) AS group_id
// 按分组聚合，获取每个连续范围的起止值
WITH group_id, MIN(num) AS range_start, MAX(num) AS range_end
// 生成最终范围字符串，单元素范围直接显示数字
RETURN COLLECT(
  CASE 
    WHEN range_start = range_end THEN toString(range_start) 
    ELSE toString(range_start) + "-" + toString(range_end) 
  END
) AS continuous_ranges

适配数据库查询场景 #

如果你的整数集合是从数据库节点/关系中聚合得到的，只需替换初始的WITH语句即可，比如从SomeNode节点的id属性获取有序集合：

MATCH (n:SomeNode)
// 确保集合有序，需在COLLECT前加ORDER BY
WITH COLLECT(n.id ORDER BY n.id) AS nums
UNWIND nums AS num
WITH 
  num,
  CASE 
    WHEN LAG(num) OVER () IS NULL OR num - LAG(num) OVER () > 1 
    THEN 1 
    ELSE 0 
  END AS is_new_group
WITH 
  num,
  SUM(is_new_group) OVER (ORDER BY num) AS group_id
WITH group_id, MIN(num) AS range_start, MAX(num) AS range_end
RETURN COLLECT(
  CASE 
    WHEN range_start = range_end THEN toString(range_start) 
    ELSE toString(range_start) + "-" + toString(range_end) 
  END
) AS continuous_ranges

内容的提问来源于stack exchange，提问作者remort