Neo4j apoc.conf

Neo4j Graph Platform Procedures & APOC
, ,
1

Hi everyone,

I'm struggling a lot trying to LOAD csv file on a daily schedule...
I've created a file with the script like in neo4j browser ( :auto load csv from ...)
then I tried to execute cypher-shell and it doesn't work - "unknown command :auto".
so, after reading, i tried "apoc.periodic.iterate('call apoc.load.csv("file....)) and also doesn't work - also tried in browser... it says:
"Failed to invoke procedure 'apoc.periodic.iterate': Caused by: java.lang.RuntimeExcpetion: import from files not enabled, please set apoc.import.fileenabled=true in your apoc.conf."

I created a file "apoc.conf" and put it in the conf folder , near neo4j.conf
in the conf file i wrote the line: "apoc.import.file.enabled=true", restart the service and doesn't help.
I don't know what to do else.
Any ideas ?
Thanks

2

@yanivts1

What version of Neo4j? I'd like to see if i can reproduce etc

3

Can you provide the cypher you are trying to execution using cypher-shell?

I just tried the following in cypher-shell and it worked.

Screen Shot 2024-06-04 at 12.35.03 PM
Screen Shot 2024-06-04 at 12.35.03 PM1568×422 46.6 KB

You would need ":auto" when executing in Neo4j Browser, but it does not look necessary when executing in cypher-shell.

1 Like
4

It's working!
Now i need to schedule it...
Thanks

1 Like
5

Hi,

When i try to load 2 million rows , the CPU goes up to 100% and everything crushes.
Any ideas?
From browser it didn't happen.

BTW, I'm using MERGE

CALL {
LOAD CSV...
MERGE (p:Person {idNumber: row.idNumber})
ON CREATE SET
p.update = datetime(),
p.name = row.name
....
}
IN TRANASCTIONS OF 500 ROWS

Thanks

6

If you are now using merge, do you have an index on Person(idNumber)?

7

yes, i have index.
Maybe the server resources is not enough?
I have 16GB RAM
Also, in the conf file i set the memory heap to 16GB and other memory things to high

8

Also, 500 rows is probably too low. Investigate 10,000 and multiples of to see if performance is better

9

Hi,

The cypher-shell worked for me and i made some progress...
Now, i want to load the data and handle errors with problematic rows.
I found some queries to do this, but again, have errors.
I'm trying in browser, not cypher-shell
so, i tried to simplify the query (without error handling) and came up with a shorter query ' just to check if it works, and it doesn't. still fail to invoke "apoc.periodic.iterate...", complaining on apoc.conf...
I'm using Neo4j 5.16, and also apoc plugins: core and extended 5.16. Also tried in Neo4j 5.20.
Here is the query:
CALL apoc.periodic.iterate(
** "CALL apoc.load.csv('file:///TEST/Persons-Full.csv',**
** {header:true}) YIELD map AS row",**
** "MERGE (s:Person {idNumber: row.idNumber})",**
{batchSize:10000, parallel:false} )

The Error:
Failed to invoke procedure apoc.periodic.iterate: Caused by: java.lang.RuntimeException: Import from files not enabled, please set apoc.import.file.enabled=true in your apoc.conf

I created a file: apoc.conf
and put 1 line:
apoc.import.file.enabled=true

What else can i try?

Thanks

10

I just verified this works. Assuming you placed the file in the database's "conf" folder, did you restart the server after adding the file?

11

Thanks,

Sure i restarted the server after putting the file in the config folder.
Anyway, it started working somehow.

1 Like