What's the issue of my cypher to remove trailing spaces?

[lingvisa]

MATCH (n) 
WITH n.Name as name
WHERE name ENDS WITH ' '
SET name = rTrim(name)
return name

I want to find nodes whose Name has trailing space and remove that space, but the command gives this error:

Type mismatch: expected Map, Node or Relationship but was String (line 4, column 12 (offset: 73))
"SET name = rTrim(name)"

The 'name' should be a string. Why does the set expects a map, node or relationship?

[tard_gabriel]

Because the SET clause expect to modify a property's value. Properties must belongs to a node or relation to be modify. In your case, name is not a property anymore but just a string, so it can't be set.

What you want to do it's keep your node to modify it's property.

MATCH (n)
SET n.name = rTrim(n.name)

Note : In this case as Neo4j will have to filter all your nodes anyway to figure out which one ends with a trailing space, so you can just apply rTrim to all of them. Might be good to double check the performance if it's a production project.

It's highly recommended to keep in lower case all your properties name.

You don't need to return anything when you WRITE something but you can still return count(*). Neo4j Desktop will provide you some hits about your query result anyway.

[lingvisa]

@tard.gabriel This is a similar query and is invalid:

MATCH (n:Product)
with n.name as Name
set n.name = replace(Name, '☆', '')
return n.name

The error message is:

Variable `n` not defined (line 3, column 5 (offset: 38))
"set n.name = replace(Name, '☆', '')"

Can't have a With statement between Match and Set? How to achieve this purpose to modify the name property?

[dana_canzano]

@lingvisa

the correct syntax is

MATCH (n:Product)
with n,n.name as Name
set n.name = replace(Name, '☆', '')
return n.name

or

MATCH (n:Product)
set n.name = replace(n.name, '☆', '')
return n.name

[lingvisa]

MATCH (n:Product)
WITH n, n.name as N
CASE WHEN N IS NOT NULL THEN replace(N,'[','') ELSE N END as newName
SET n.name = newName
RETURN count(*)

But I received this error:

Invalid input 'S': expected 'l/L' (line 3, column 3 (offset: 70))
"CASE WHEN ING IS NOT NULL THEN replace(ING,'[','') ELSE ING END as result"
   ^

@dana.canzano Another similar one. Why is the syntax wrong?

[dana_canzano · Accepted solution]

@lingvisa

Neo4j version? if this is a bug I'd like to test/reproduce against the same version you have encountered with

But also

MATCH (n:Product)
WITH n, CASE n.name when NOT NULL THEN replace(n.name,'[','') ELSE n.name END as newName
SET n.name = newName
RETURN count(*)

should accomplish what is desired

[lingvisa]

I initially tested in 4.0.4, then tested in 4.2.6, and the message is more informatvie as below. Perhaps it doesn't like "CASE" starts with a new cypher statement:

Invalid input 'CASE': expected 
  <EOF>
  "RETURN"
  "CREATE"
  "DELETE"
  "SET"
  "REMOVE"
  "DETACH"
  "MATCH"
  "WITH"
  "UNWIND"
  "USE"
  "CALL"
  "LOAD"
  "FROM"
  "FOREACH"
  "WHERE"
  "MERGE"
  "OPTIONAL"
  "ORDER"
  "SKIP"
  "LIMIT"
  "UNION"
  "," (line 3, column 1 (offset: 37))
"CASE N when NOT NULL THEN replace(n.name,'[','') ELSE n.name END as newName"
 ^

[dana_canzano]

@lingvisa

Thanks for the version detail.

The cypher I provided did not fail. Is your last post an error from the cypher I had provided ? If not what cypher is now being submitted

[lingvisa]

Your cypher is good. The last error message was produced from 4.2.6 version, and I think it's more informative. Probably it doesn't like CASE starts a new statement.