How to optimize a query with a large number of hops

[LianxinGao]

Like cypher:

MATCH (:Country)<-[:IS_PART_OF]-(:City)<-[:IS_LOCATED_IN]-(:Person)<-[:HAS_MEMBER]-(:Forum)-[:CONTAINER_OF]->(:Post)<-[:REPLY_OF]-(:Comment)-[:HAS_TAG]->(:Tag)-[:HAS_TYPE]->(:TagClass)
RETURN count(*) AS count

This query has 7 hops, if each hop has 100 data, then total data will be (100)^7， it's a huge amount of data. Besides, currently, we will deserialize all the data's properties, so OOM will happen.....

[LianxinGao]

My idea is:

1. optimize the query in lynx, if we only need to get the count number, then when move to the next hop, we can drop the previous data?
2. optimize the data structure, use lazy node and lazy relationship( do not deserialize the properties, just use the id...)

@chunyang-wen @zyang-tudb @xuanying2020 @terrytangyuan do your have any good ideas？