MironAtHome commented on issue #1974: URL: https://github.com/apache/age/issues/1974#issuecomment-2230034348
> I restarted the server and tried again with `SET LOCAL parallel_leader_participation = off;` and the results are same. So, it clearly works. > > If you give any direction query runs much much faster and yes seems like these two produce idential output. Query 1: https://explain.dalibo.com/plan/gc622ba38g91afb1 > > Query 2: https://explain.dalibo.com/plan/14bd391394c9758c After studying closely behavior of operators ()-[]-() and ()-[]->() it is evident, that data will be presented as following: If, for example, one is to create two vertices v:Person {name:'Alice'} v:Person {name:'Bob'} and follows by creating two edges Alice-[:knows]->Bob and Alice<-[:knows]-Bob with the id's assigned as following: Alice-id:1 Bob-id:2 Edge :knows for the relationship Alice-[:knows]->Bob id 3 Edge :knows for the relationship Alice<-[:knows]-Bob id 4 Query Match (v1)-[:knows]-(v2) Will return the following combination of id's Row #1: 1,3,2 Row #2: 1,4,2 Row #3: 2,3,1 Row #4: 2,4,1 However, if we were to present not id's ( which is really an system values used by machine internals ) and produce just the distinct combination of names Match (v1)-[:knows]-(v2) WITH v1.name as Name1, v2.name as Name2 RETURN DISTINCT Name1, Name2 ORDER BY Name1 ASC, Name2 ASC We will get back Alice, Bob Bob, Alice Now, if we were to issue query (v1)-[:knows]->(v2) For the id producing query we will get only half the rows Row #1: 1,3,2 Row #2: 2,4,1 However, for the name producing query, there would be no difference. In other words, (v1)-[:knows]-(v2) Behaves as (v1)-[:knows]->(v2) UNION (v1)<-[:knows]-(v2) To summarise, looking at the last query just one line above, it's 2x the number of rows stored, with the same relationship listed twice, except switching the sides of the data. In case producing rows once with relationship going left to right is acceptable, query can be expressed as (v1)-[:<edge>]->(v2) without losing resolution. And finally, if performance of query is essential, it could be of use to produce various permutations of data using client? Hope this helps. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: dev-unsubscr...@age.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
