CODEWITHSUNDEEP
neo4j
bjlevine
bjlevine

Reputation: 984

Difference between a single (OPTIONAL) MATCH clause and multiple

What is the difference between

OPTIONAL MATCH clauseA, clauseB

and 

OPTIONAL MATCH clauseA
OPTIONAL MATCH clauseB

I get different behavior depending on which form I use.

For example:

START n=node(111)
OPTIONAL MATCH n<-[links_n_in]-(n_from),n-[links_n_out]->(n_to)
RETURN n,COLLECT(n_from) AS n_from,COLLECT(links_n_in) AS links_n_in,COLLECT(n_to) AS n_to,COLLECT(links_n_out) AS links_n_out

which is designed to return a node; it's incoming relationships and from nodes; it's outgoing relationship and to nodes.

I have a test graph consisting of Node 111 which has 4 outgoing relationships each of which points to the same Node (I have other test cases in which 111 points to different Nodes). Executing the query as above returns only Node 111 in column 'n'. The columns for 'n_from', 'links_n_in', 'n_to', 'links_n_out' are empty.

If I modify the query to:

START n=node(111)
OPTIONAL MATCH n<-[links_n_in]-(n_from)
OPTIONAL MATCH n-[links_n_out]->(n_to)
RETURN n,COLLECT(n_from) AS n_from,COLLECT(links_n_in) AS links_n_in,COLLECT(n_to) AS n_to,COLLECT(links_n_out) AS links_n_out

then the n_to and link_n_out columns are populated as expected.

Upvotes: 2

Views: 599

Answers (1)

Eve Freeman
Eve Freeman

Reputation: 33155

The first form treats it as a single extended pattern that must match entirely.

The second form treats them as distinct optional patterns, and can match the two separately.

So your results make sense, when you think about what it's doing--if the whole OPTIONAL MATCH pattern isn't found, it doesn't match any of the OPTIONAL MATCH pattern.

Upvotes: 5

Related Questions