汇总后计算

Question

我是Neo4j的新手，我觉得自己陷入了一些简单的操作中，这些操作我会使用子查询在常规SQL中解决。

如何减去两个结果行？我将结果分组，我想返回它们之间的差异：

MATCH (seguidores:RelevantTwitterUser {location:"Madrid"})-[:FOLLOWS]->(seguidos:RelevantTwitterUser {location:"Barcelona"}) 
WITH COLLECT({origen:seguidores.location, user:seguidores.userId}) AS ROWS
MATCH (seguidores:RelevantTwitterUser {location:"Barcelona"})-[:FOLLOWS]->(seguidos:RelevantTwitterUser {location:"Madrid"}) 
WITH ROWS + COLLECT({origen:seguidores.location, user:seguidores.userId}) AS allRows
UNWIND allRows AS ROW
RETURN ROW.origen, COUNT(ROW.user)

输出：

Answer 1

使用aggregating functions

WITH "Madrid" AS loc1, "Barcelona" AS loc2
MATCH (:RelevantTwitterUser{location:loc1})-[:FOLLOWS]->(:RelevantTwitterUser{location:loc2})
WITH loc1, loc2, COUNT(*) AS count1
MATCH (:RelevantTwitterUser{location:loc2})-[:FOLLOWS]->(:RelevantTwitterUser{location:loc1})
WITH loc1, loc2, count1, COUNT(*) AS count2
RETURN loc1, count1, loc2, count2, count1 - count2 AS diff

如果您想了解此查询的工作原理，并应该阅读aggregating functions（如COUNT）上的文档，并在需要修改此查询时避免计数错误。了解grouping keys（例如最后一个locs子句中的count和WITH）如何影响聚合函数的行为尤其重要。

使用SIZE() on pattern expressions

WITH "Madrid" AS loc1, "Barcelona" AS loc2
WITH loc1, loc2,
  SIZE((:RelevantTwitterUser{location:loc1})-[:FOLLOWS]->(:RelevantTwitterUser{location:loc2})) AS count1,
  SIZE((:RelevantTwitterUser{location:loc2})-[:FOLLOWS]->(:RelevantTwitterUser{location:loc1})) AS count2
RETURN loc1, count1, loc2, count2, count1 - count2 AS diff