Neo4j: Sequence of Events as Nodes Not working - neo4j

Question

I am new to cypher query syntax and tried different types of syntax/relationship to build sequence graph. My data contains group_id and within each group_id a code occurs based on the 'number'. Lowest number is the first sequence and highest number is the last sequence per group id. I am able to load the data from csv and create nodes with properties, however it is not letting me convert to numerical sequence for 'code' nodes. I am reading/referencing this article : this tutorial. Is there special cypher syntax to use to achieve this result?
Sample Data:
group_id,code,date,number
123,abc,2/18/21,4
123,def,11/11/20,3
123,ghi,11/10/20,2
123,jkl,10/1/20,1
456,gtg,11/28/20,5
456,abc,10/30/20,4
456,def,10/5/20,3
456,jkl,10/1/20,2
456,uuu,10/1/20,1
My Code to load data:
LOAD CSV WITH HEADERS FROM "file:///sample2.csv" AS row
WITH row
WHERE row.group_id IS NOT NULL
MERGE (g:group_id {group_id: row.group_id});
LOAD CSV WITH HEADERS FROM "file:///sample2.csv" AS row
WITH row
WHERE row.code IS NOT NULL
MERGE (c:code {code: row.code})
ON CREATE SET c.number = row.number,
c.date = row.date;
Here is what I have tried:
// Building relationship
LOAD CSV WITH HEADERS FROM "file:///sample2.csv" AS row
WITH row
MATCH (g:group_id {group_id: row.group_id})
MATCH (c:code {code: row.code})
MERGE (g)-[:GROUPS]->(c) // Connects ALL codes to group id, but how to connect to 'code' and 'number' sequentially?
MERGE (c:{code: row.number})-[:NEXT]->(c) // Neo.ClientError.Statement.SyntaxError
I have gotten result:
I am trying to get this.

Answer 1

This will be a two step process. First the initial loading of the data as you have outlined. Then an enhancement in which you create the NEXT relationships. We do this in healthcare analytics of patient journeys or trajectories. By analogy, your yellow nodes might be a patient and the blue one an encounter. So each patient has a sequence of encounters.
You can query and sort by the date or other ordering variable. For example, collect a sorted list of encounters:
match (e:encounter) with e order by e.enc_date with e.subjectId as sid,collect(distinct e.enc_date) as eo return sid,size(eo) as ct,eo
I used this in some python code to then iterate through the collection to create the enc_seq edge, equivalent to your NEXT:
> dfeo = Neo4jLib.CypherToPandas("match (e:encounter) with e order by e.enc_date with e.subjectId as sid,collect(distinct e.enc_date) as eo return sid,size(eo) as ct,eo",'ppmi')
csv = dfeo.to_csv(index=False).split('\n')
cts=0
sw = open("c:\\temp\\error.txt","a")
for i in range(1,len(dfeo)):
cc = csv[i].split(',')
for j in range(0,int(cc[1])-1):
try:
q= "match (e1:encounter{subjectId:" + str(dfeo['sid'][i]) + ",enc_date:date('" + str(dfeo['eo'][i][j]) + "')}) match (e2:encounter{subjectId:" + str(dfeo['sid'][i]) + ",enc_date:date('" + str(dfeo['eo'][i][j+1]) + "')}) merge (e1)-[r:enc_seq{subjectId:" + str(dfeo['sid'][i]) + ", seqCt:" + str(j) + "}]-(e2)"
Neo4jLib.CypherNoReturn(q,'ppmi')
except:
cts = cts + 1
sw.write(str(i) + ':' + str(j) + "\n"+ q + "\n")
print("exceptions: " + str(cts))
sw.flush()
sw.close()
You can probably do this within a cypher query using a WITH (each row) followed by a CALL to a function similar to my python code. For my purposes it was more convenient to use python.