I have a graph of members and the items that they've looked at.
This data is going to be used to recommend items based on items that similar members have looked at. I'd like to sort the items based on how similar the colors of the items are. The colors are stored on the items in an array (["red", "blue", "green"]). Is there any way in cypher to compare arrays to see how many elements they have in common?
Given two nodes, n and m, that look something like:
CREATE ({id: 1, color: ["red", "blue", "green", "yellow"]})
CREATE ({id: 2, color: ["red", "blue", "green", "white"]})
You can do something like this:
MATCH n, m
WHERE n.id = 1 AND m.id = 2
RETURN length(FILTER(x in n.color WHERE x in m.color))
The FILTER
function iterates through the n.color array, binding the current value to x
(arbitrarily chosen by me, could be different). The predicate (x in m.color
) is checked for each x
value, and if it evaluates to true, that element is pushed into the new array that FILTER
returns. You can leave it at that to see the intersection of the two arrays (red, blue, and green in this case), or wrap it in the length
function to see the number of colors shared between the two nodes (3 in this case).
Check out the full FILTER docs here: http://docs.neo4j.org/chunked/milestone/query-functions-collection.html#functions-filter
The solution given above is good. But in the new version(4.0), they have made some changes. The Filter() method is deprecated and list() is
Restricted to only work on paths
in Version 4.0
They replaced the filter with List comprehension so instead of using the above code
MATCH n, m
WHERE n.id = 1 AND m.id = 2
RETURN length(FILTER(x in n.color WHERE x in m.color))
I recommend you to use this one.
MATCH n, m
WHERE n.id = 1 AND m.id = 2
RETURN size([x in n.color WHERE x in m.color])
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With