Hi @keith_ivanchan
A very clever colleague of mine, @jfgirouard, created an app that uploads nodes and relations to a local NEO4J instance using the API. Each of the nodes is a json statement that is set as the payload of Dynamic Input Post command
{ "statements" : [ { "statement" : "MERGE (n:SourceEntity { name:\"EntityName\"}) RETURN n" } ] }
Once the nodes are uploaded, the relations are sent
{ "statements" : [ { "statement" : "MATCH (a:Table),(b:Table) WHERE a.name = \"EntityName1\" AND b.name = \"EntityName2\" CREATE (a)-[r:REFERENCES {} ]->(b) RETURN r" } ] }
The JSON statements are built on the fly from information pulled from one of our databases. The strings are built from the inside out, first building out the Cypher strings and then wrapping them in JSON. The heavy lifting is done by a series of formula tools that update EntityName1 and EntityName2
The process for creating Cypher files should be similar, minus the JSON wrapper part
MERGE (n)
ON CREATE SET n.prop = 0
MERGE (a:A)-[:T]-(b:B)
ON CREATE SET a.name = 'me'
ON MATCH SET b.name = 'you'
RETURN a.prop
This statement could be built using a formula like
"MERGE (a:A)-[:T]-(b:B)
ON CREATE SET a.name = '" + [A] +"'
ON MATCH SET b.name = '" + [B] +"'"
Of course on a large scale, you'd probably want to have the boiler plate set as a series of templates like this(assuming all Merge statement have the same format)
MERGE (a:A)-[:T]-(b:B)
ON CREATE SET a.name = '__ANAME__'
ON MATCH SET b.name = '__BNAME__'
and use some mechanism to replace __ANAME__ and __BNAME__ at run time
Of course, the hard part will be digesting the Cypher manual:)
Good Luck
Dan
