search

vesoft-inc / nebula-python

Client API of Nebula Graph in Python
191 stars 75 forks source link

Querying NebulaGraph using the Nebula Python API is significantly slower than using the Nebula Console. #266

Closed neojoe closed 11 months ago

neojoe commented 1 year ago

我在服务器上用docker的方式安装nebulagraph 服务器配置：cpu 16v ，内存128g，150G ESSD云盘 执行相同的查询逻辑 nebula onsole：MATCH p = allShortestPaths((a:point{name:"13644:5"})-[e*..7]-(b:point{name:"5119:1"})) RETURN p; 查询时间是700ms左右 image 而用nebula python api 执行相同的逻辑，花费65s image 请问是什么原因？

neojoe commented 1 year ago

java api 也是同样的问题

wey-gu commented 1 year ago

Ping @Aiee @Nicole00 is this related to serialization time, or?

Nicole00 commented 12 months ago

你可以看下console中执行结果会有两个时间，700ms 是执行结果ResultSet中的latency时间，后面那个是客户端执行的总体时间。

如果要对比其他客户端，可以通过同时对比resultSet.getLatency 或者 同时对比 客户端执行总时间 来确定各个客户端的性能差异。

QingZ11 commented 11 months ago

@neojoe hi, if your problem is solved, would you like to close the issue? Thanks a lot for your contribution.