所有简单（循环）路径查询 returns 数据过多 -> 堆溢出

Question

我是 Neo4j 的新手，使用带有 Javascript 驱动程序的 Neo4j 桌面版。我的图有大约 30.000 个节点和 40.000 个边。最终的方法是获取图中的所有简单循环，但我在尝试执行以下密码查询时遇到堆溢出：

let res = await session.run('MATCH p=(n)-[*2..4]-(n) RETURN nodes(p)')

这当然是因为查询尝试 return 找到的所有简单循环，这对我的堆来说太过分了，不幸的是，增加 node.js 的内存不是一个选项。

那么有什么方法可以绕过这个问题并从图中获取所有简单循环吗？

这是我得到的堆溢出错误：

/Users/paulus/.nvm/versions/node/v15.10.0/bin/node /Users/paulus/routeplanner/RouteFinder.js

<--- Last few GCs --->

[39630:0x104e00000]   227813 ms: Scavenge (reduce) 2033.5 (2077.9) -> 2032.1 (2072.1) MB, 113.5 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 
[39630:0x104e00000]   228171 ms: Scavenge (reduce) 2036.5 (2072.6) -> 2035.5 (2075.6) MB, 28.6 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 
[39630:0x104e00000]   228558 ms: Scavenge (reduce) 2039.1 (2075.6) -> 2038.5 (2080.4) MB, 28.2 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 


<--- JS stacktrace --->


<--- Last few GCs --->

[39630:0x104e00000]   227813 ms: Scavenge (reduce) 2033.5 (2077.9) -> 2032.1 (2072.1) MB, 113.5 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 
[39630:0x104e00000]   228171 ms: Scavenge (reduce) 2036.5 (2072.6) -> 2035.5 (2075.6) MB, 28.6 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 
[39630:0x104e00000]   228558 ms: Scavenge (reduce) 2039.1 (2075.6) -> 2038.5 (2080.4) MB, 28.2 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 


<--- JS stacktrace --->

FATAL ERROR: MarkCompactCollector: young object promotion failed Allocation failed - JavaScript heap out of memory
FATAL ERROR: MarkCompactCollector: young object promotion failed Allocation failed - JavaScript heap out of memory

<--- Last few GCs --->

[39630:0x104e00000]   227813 ms: Scavenge (reduce) 2033.5 (2077.9) -> 2032.1 (2072.1) MB, 113.5 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 
[39630:0x104e00000]   228171 ms: Scavenge (reduce) 2036.5 (2072.6) -> 2035.5 (2075.6) MB, 28.6 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 
[39630:0x104e00000]   228558 ms: Scavenge (reduce) 2039.1 (2075.6) -> 2038.5 (2080.4) MB, 28.2 / 0.0 ms  (average mu = 0.577, current mu = 0.321) allocation failure 


<--- JS stacktrace --->

FATAL ERROR: MarkCompactCollector: young object promotion failed Allocation failed - JavaScript heap out of memory

Process finished with exit code 139 (interrupted by signal 11: SIGSEGV)

Answer 1

如果不了解图表的数据模型，很难给出准确的解决方案，但使用 skip/limit 获取更小批量的数据可能会有所帮助。

这将获得前 100 个模式：

MATCH p=(n)-[*2..4]-(n) RETURN nodes(p) SKIP 0 LIMIT 100

接下来的 100 个：

MATCH p=(n)-[*2..4]-(n) RETURN nodes(p) SKIP 100 LIMIT 100

所有简单（循环）路径查询 returns 数据过多 -> 堆溢出

All simple (cyclic) paths query returns too much data -> heap overflow

node.js

neo4j

cypher

neo4j-driver

neo4j-desktop