Fixes #3751: mapParallel2 with rebind error

[vga91]

Fixes #3751

The error seems to be due to the rebind introduced in 5.9 (https://github.com/neo4j-contrib/neo4j-apoc-procedures/blob/dev/extended/src/main/java/apoc/cypher/CypherExtended.java# L248).

Sporadically the rebind throws (under the hood), an exception like:

org.neo4j.exceptions.UnderlyingStorageException: Access to record Node[20,used=false,created=false,rel=-1,prop=-1,labels=Inline(0x0:[]),light,fixedRefs=false] went out of bounds of the page. The record size is 15 bytes, and the access was at offset 300 bytes into page 0, and the pages have a capacity of 8192 bytes. The mapped store file in question is /target/test data/neo4j/data/databases/neo4j/neostore.nodestore.db

or:

org.neo4j.graphdb.NotFoundException: Node 4:cce42f14-8960-44a9-b61f-5cdf5d7f0bc9:420 not found.

Resolution

Discussing it here,
the problem seems to be due to the rebind, which instead of using the @Context Transaction tx as a transaction should use the one coming from db.beginTx().
In this way, it works correctly.

The same method is also used by other procedures, but these also work with @Context Transaction tx, maybe because they don't create parallel partitions.
I created tests to be sure, so I would leave it like this so as not to cause breaking-change.
Also because is not possible to change this one, since it would give an error A query with 'CALL { ... } IN TRANSACTIONS' can only be executed in an implicit transaction