CSV Import

Neo4j Graph Platform Import / Export
1

I am planning to load train schedule and stations into Neo4j from CSV.

Source Data

TrainNo TrainName   SEQ   StationCode   Arrival Departure Distance
1           TN_1          1      S1           8      9       0
1           TN_1          2      S2           10     11     10
1           TN_1          3      S3           12     1      15 
1           TN_1          4      S4           3      4       15
2           TN_2          1      S1         
2           TN_2          2      S2         
2           TN_2          3      S5         
2           TN_2          4      S6         
2           TN_2          5      S7         
2           TN_2          6      S8

I need to build nodes and relationship like this

S1--(TrainNo,TrainName,SEQ,Arrival,Depature,Distance)--S2--(TrainNo,TrainName,SEQ,Arrival,Depature,Distance)--S3--(TrainNo,TrainName,SEQ,Arrival,Depature,Distance)-S4

Basically, the TrainNo, TrainName,Seq, Arrival, Depature and Distance will be on the relationships, and the same relationships will form a route between the stations.

Neo4j - 3.5

2

Hi There, maybe using APOC's?

USING PERIODIC COMMIT 3000
LOAD CSV WITH HEADERS FROM "file:///trains.csv" AS csvLine
FIELDTERMINATOR ';'MERGE (start:Station{ stationName: csvLine.startStation})
MERGE (destination:Station{ stationName: csvLine.nextStation})
MERGE (line:Line{ lineName: csvLine.lineName})
WITH start, destination,csvLine.lineNameas lineName
CALL apoc.create.relationship(start, lineName, NULL, destination) YIELD rel return count(*)

3

hi @jhonathanmarolo, there is no column like startStation and nextStation. The station are in sequence by row by row.

4

@david_allen @michael.hunger

5

This solution works for your data displayed.

!. Created the station nodes:

USING PERIODIC COMMIT
LOAD CSV WITH HEADERS FROM 'file:/am3.csv' AS line
WITH line

MERGE (s:Station {Station:line.Code, SEQ: toInteger(line.SEQ), TrainName:line.Train, TrainNo:toInteger(line.TrainNum)})
;

  1. Create relationships

USING PERIODIC COMMIT
LOAD CSV WITH HEADERS FROM 'file:/am3.csv' AS line
WITH line

FOREACH(ignoreMe IN CASE WHEN toInteger(line.SEQ) > 1 THEN [1] ELSE END |

MERGE (s:Station {TrainName:line.Train, SEQ: (toInteger(line.SEQ)-1)})
MERGE (s1:Station {TrainName:line.Train, SEQ: toInteger(line.SEQ)})
MERGE (s)-[:NEXT]->(s1)

)
;

Result:
am2

A robust solution is to have a column that stores the previous station id or name. Then all nodes and relationships can be created in one run. One option that I tried successfully is to import dataset into SQL server and use Windows functions to achive the desired result.

1 Like