Neo4j - 查询：将不同的查询结果组合在一个具有特殊顺序的集合中

Question

我有这样的查询

START a = node:lucene_index('name: "actor1"')
MATCH (a:Actor)--(b:Movies)
WHERE b.director = 'director1'
WITH b LIMIT 4
RETURN 'Movies' AS resultType, collect (DISTINCT b) as resultList

UNION ALL 

START a = node:lucene_index('name: "actor1"')
MATCH (a:Actor)--(b:Movies)
WHERE b.director = 'director2'
WITH b LIMIT 4
RETURN 'Movies' AS resultType, collect (DISTINCT b) as resultList

UNION ALL 

START a = node:lucene_index('name: "actor1"')
MATCH (a:Actor)-[r]-(b:Movies)
WITH b
MATCH (b)-[r2]-(b2:Actor)
WITH b, count(r2) as relations
ORDER BY toInt(relations) DESC
LIMIT 2
RETURN 'Movies' AS resultType, collect (DISTINCT b) as resultList

是否可以将这些查询合并为一个？ 重要的是，resultList按照查询

中所示的顺序包含电影

• 1 - 4 of director1
• 5 - 8 of director2
• 9 - 10与演员关系最多

谢谢！

Answer 1

您可能需要查看post-union processing上的Neo4j知识库文章。这提供了一些方法来执行您的工会，但之后能够执行排序操作。

当然，你需要一些东西来解决这个问题。我会为每个（例如order）添加第三个字段，并为第一个子查询显式设置为1，第二个子查询显式设置为2，第三个子查询显式设置为3。然后在联合后处理时对该字段进行排序。