Cypher grammer, syntax, and documentation -- List, Collect, avg, etc are impenetrable to me

Neo4j Graph Platform Cypher
2

Half of Cypher's magic is in transformation between different groupings of data. This really boils down to one big matrix, which can be thought of like a table.

  • Any operation will act on one "row" at a time, until it has done all rows.
  • Any variable or property in that row can be a list.
  • Aggregation functions, and things like COLLECT take multiple rows, and turn them into one variable in one row.
  • An expression is a part of a Cypher command which "chooses" specific variables from a row.
  • WITH chooses specific variables to pull from all rows, and use in following commands.
  • UNWIND turns a list (or collection) into rows.

So, to answer your question, if you're working with a list, or collection, you have two choises:

  • Use UNWIND to turn the list into rows, then you can apply aggregation functions like avg().
  • Use list functions like REDUCE() to collapse the list into a single variable.

UNWIND

WITH [1, 2, 3, 4, 5] AS list1
UNWIND list1 AS vals
RETURN avg(vals) AS average

In this example, we're turning list1 into a bunch of "rows," where each entry in the list is now referenced by the vals variable. The expressions passed to the avg function is selecting all instances of vals as input to avg.

REDUCE

WITH [1, 2, 3, 4, 5] AS list1
RETURN REDUCE(sum = 0, val IN list1 | sum + val) / SIZE(list1) AS average

Instead of turning the list into a set of rows, we're stepping through all entries in the list, doing some math, and spitting out the result. Essentially creating our own avg function for operating on lists.

Apoc wins everything

No matter what you're doing, APOC probably has a clean, fast, tool that will perfectly fit your need. This case included. Enter, APOC Collection Functions

WITH [1,2,3,4,5] AS list1
RETURN apoc.coll.avg(list1) AS average

See Also: cypher - How to use average function in neo4j with collection - Stack Overflow