问题
I have created fresh dataset to explain my desired result. and here is the link
Or you can trigger this command using cypher.
create
(_6 {UserName:\"dhansukh\", UserProfileID:\'1000\', EMailID:\'f@xyz.com\'}),
(_5 {UserName:\"dhruv\", UserProfileID:\'516\', EMailID:\'e@xyz.com\'}),
(_4 {UserName:\"dharmistha\", UserProfileID:\'5262\', EMailID:\'d@xyz.com\'}),
(_3 {UserName:\"dinesh\", UserProfileID:\'995\', EMailID:\'c@xyz.com\'}),
(_2 {UserName:\"dharmesh\", UserProfileID:\'502\', EMailID:\'b@xyz.com\'}),
(_1 {UserName:\"manish\", UserProfileID:\'1\', EMailID:\'a@xyz.com\'}),
_1-[:friends {ApprovalStatus: 1} ]->_2,
_1-[:friends {ApprovalStatus: 1} ]->_3,
_1-[:friends {ApprovalStatus: 2} ]->_5,
_2-[:friends {ApprovalStatus: 1} ]->_3,
_2-[:friends {ApprovalStatus: 1} ]->_5,
_3-[:friends {ApprovalStatus: 1} ]->_4
Now I am trying following query, but it is not given me my expected result.
START me=node:node_auto_index(UserProfileID = \'1\'), other=node(*)
MATCH pMutualFriends=me-[r?:friends]-mf-[r1:friends]-other
WHERE other.UserName =~ \'(?i)dh.*\' AND other.UserProfileID <> 1
RETURN other.UserProfileID, other.UserName, r.ApprovalStatus, COUNT(pMutualFriends) AS mutualCount
In the above result set, I have get duplicate records, (due to ApprovalStatus), If I remove ? from relationship, it just shows linked node only, but I want all nodes started with \'dh\'. node 6 also missing, don\'t know why? mutual count also showing wrong in some case. Only that node should be consider in mutual count which is has ApprovalStatus = 1. like login node (eg. node 1) and search nodes both have relationship\'s property ApprovalStatus = 1.
EDIT : My expected result set :
UserProfileID UserName ApprovalStatus MutualCount
------------- -------- -------------- -----------
502 dharmesh 1 2 (node 3 & 5 )
516 dhruv 2 1 (node 2)
5262 dharmistha null 1 (node 3)
1000 dhansukh null 0
EDIT : I am updating image for clear understanding.
I am suffering from this issue last 20-25 days, and not getting proper solution, I don\'t know where is the problem. I have already post this problem many times on stackoverflow. here is the link, and this and this and many more.
回答1:
As I said in my comment, trying to query for connected and disconnected nodes at the same time doesn't seem to be a good idea.
If you want only connected nodes, try the following query :
START me=node:node_auto_index(UserName = 'manish')
MATCH me-[:friends]-mf-[:friends]-other, me-[r?]-other
WHERE other.UserName! =~ '(?i)dh.*'
RETURN DISTINCT ID(other), r.ApprovalStatus AS status, count(mf) AS mutual, ID(me)
ORDER BY mutual DESC , status ASC
Please note that I had to add another pattern in the match clause, because your approval status is between (me) and (other), and not between (me) and (mutual friend), which is what you were doing!
This will return the first 3 lines of your expected answer and ignores dhansukh.
回答2:
What about this? http://console-test.neo4j.org/?id=8lloq1
START me=node:node_auto_index(UserProfileID = '1'), other=node(*)
MATCH pMutualFriends=me-[r?:friends]-mf-[r1?:friends]-other
WHERE other.UserName! =~ '(?i)dh.*' AND other.UserProfileID? <> 1 AND r.ApprovalStatus=1
RETURN DISTINCT (other.UserProfileID?),ID(other),me.EMailID?, other.EMailID?, other.UserName?, r.ApprovalStatus, COUNT(pMutualFriends) AS mutualCount
来源:https://stackoverflow.com/questions/17947903/search-list-with-mutual-count-2nd-try