if I delete a top node containing nested (connecte...
# orm-helpz
zonofthor
09/04/2018, 9:00 AMif I delete a top node containing nested (connected) nodes - should I delete the sub-nodes so that they aren't left as "undeserted" ? Are they cleaned up automatically by some health-check script? I am trying to not have to define relationship in SDL, I know using cascade will take care of this... but is it really needed?
a
Arnab
09/04/2018, 9:08 AMSo you basically have lots of
Copy code
type A {
b: B @relation(name: "AtoB", onDelete: CASCADE)
otherField: String
...
}
type B {
...
}ABz
zonofthor
09/04/2018, 9:09 AMyes but I want to be lazy and not define the @relation (...)
zonofthor
09/04/2018, 9:10 AMas B is only related to A I was hoping it would be deleted automatically somehow, by some maintenance routine maybe...
a
Arnab
09/04/2018, 9:10 AMwell if you don't define cascade, how would prisma know that nodes of type `B`should be deleted when nodes of type
ABArnab
09/04/2018, 9:12 AMExample: Due to GDPR, a lot of sensitive data must be removed from a
PersonpersonaliaPerson objectz
zonofthor
09/04/2018, 9:12 AMit couldn't - but they would not be connected to any node at all
a
Arnab
09/04/2018, 9:13 AMThat's my point. I'd be very surprised if Prisma suddenly started making decisions like this to data that is not connected
z
zonofthor
09/04/2018, 9:13 AMhmm... I got enlightened... I was under the impression that connected nodes were different somehow - thanks for clarifying!
a
Arnab
09/04/2018, 9:14 AMWhat do you mean?
z
zonofthor
09/04/2018, 9:14 AMI'll define all relations in schema from now on!
zonofthor
09/04/2018, 9:15 AMdoesn't matter, just my misunderstanding - but thx again for clarifying!
a
Arnab
09/04/2018, 9:15 AMIf you want prisma to have some kind of cleanup routine that cleans unreachable data, then your very first datatype would automatically be cleaned up 😛
z
zonofthor
09/04/2018, 9:15 AMit all started when I read that relationship names don't have to be defined in SDL
zonofthor
09/04/2018, 9:16 AMso I started removing the @relationship declarations because I thought it was overhead
a
Arnab
09/04/2018, 9:16 AMAh
Arnab
09/04/2018, 9:17 AMI can see that. You assumed that on-cascade-delete was the default?
z
zonofthor
09/04/2018, 9:17 AMexactly!
zonofthor
09/04/2018, 9:18 AMalso bothers my that to defined delete action I have to create a name for the relation
@relation(name: "Comments", onDelete: CASCADE)zonofthor
09/04/2018, 9:19 AMmaintaining these names was becoming a tedious task for a schema in creation
a
Arnab
09/04/2018, 9:21 AMHmm. I think I'm confused by the docs as well here
Arnab
09/04/2018, 9:21 AMThis argument is only required if relations are ambiguous. Note that the name argument is required every time you're using the @relation directive.
z
zonofthor
09/04/2018, 9:31 AMagree - "you can skip it but still you need it"
6 Views