Help understanding collection indexing

[usuallycwdillon]

I'm trying to write a query that iterates over a collection of nodes by their index (so that I can iterate two lists in tandem). I'm not getting expected results.

For example, if I have 10 nodes with label "Node" in my graph and I query MATCH (n:Node) WITH n, collect(n) AS myNodes RETURN size(myNodes) I expect to get one result

size(myNodes)
10

Instead, I get 10 results

size(myNodes)
1
1
1
...

I expected from the answer to this SO question that I should be able to iterate over the results of a collect() function by the index.

[elaine_rosenber · Accepted solution]

Change:

MATCH (n:Node) WITH n, collect(n) AS myNodes RETURN size(myNodes)

To:

MATCH (n:Node) WITH collect(n) AS myNodes RETURN size(myNodes)

Elaine

[usuallycwdillon]

Thank you, Elaine.

Just so that I understand this better... querying WITH n, collect(n) is creating a list of the collection (with just one element), where WITH collect(n) creates the list of the results. Is that what's going on?

[andrew_bowman]

Aggregation in Cypher is with respect to the set of non-aggregation variables when you perform the aggregation, which forms the grouping key.

In your query, WITH n, collect(n), this means collecting n with respect to each n. n is the grouping key. You'll get an n per row, and the collection of just itself in the list.

When you do WITH collect(n), there are no other variables besides the aggregation, so it collects over everything, and you get a single row with the entire collected set of nodes.