On a sharded collection, if you try to create a custom index (e.g. hash index) with uniqueness, that doesn't include the sharding key (which by default is the autogenerated _key) then it will fail with an error like:

Index creation failed: shard key '_key' must be present in unique index

There's zero documentation that I could find around this, and it's fairly fundamental, since you can't

When checking if a document exists in a collection, I use the function documentCollection.documentExists. For example:

const cities = db.collection("City")
const city_exists = await cities.documentExists("Country/1234")
console.log("city_exists = ", city_exists)

But instead of looking only in the specific "City" collection, it finds documents from other collections (like "Count

I've been working on a velocypack parser (in Rust), and ran into some troubles parsing some output from ArangoDB:

0x0c 0x73 0x01 0x12 0x00 0x00 0x00 0x00 0x00 0x4a 0x61 0x75 0x74 0x68 0x6f 0x72 0x69 0x7a 0x65 0x64 0x1a 0x44 0x75 0x73 0x65 0x72 0x18 0x4b 0x69 0x73 0x41 0x64 0x6d 0x69 0x6e 0x55 0x73 0x65 0x72 0x1a 0x48 0x64 0x61 0x74 0x61 0x62 0x61 0x73 0x65 0x47 0x5f 0x73 0x79 0x73 0x74 0x6

According to https://www.arangodb.com/docs/stable/transactions-locking-and-isolation.html it is possible to pass collections: { exclusive: [...] } to executeTransaction, though no examples include it.

public function update(Document $document, array $options = [])
{
$documentId = $this->getDocumentId($document);
return $this->updateById($document, $documentId, $document, $options);
}

It seems the first arg to call to updateById should be $collection, and the update function does not accept $collection. Am I missing something?

I get a com.arangodb.ArangoDBException: Response: 404, Error: 1202 - document not found exception, when I call ArangoVertecCollectionImpl#insertVertex().

Isn't that a bug? I'm not sure?

Now that ArangoDB supports in database document validation you could consider to delete the client validation code or at least deprecate the functionality.

Doing the setup on GKE would be probably way easier if users would be pointed to the document describing the required steps for GKE.

Is there implementation of AQL's DOCUMENT function which returns document by id?

Basically every time you're using graph operations such as GRAPH_NEIGHBORS, GRAPH_COMMON_NEIGHBORS etc. you're getting list of handles back and in order to get the actual documents you need to perform join on id with vertex collections whereas AQL natively supports document function which returns document by handle

Hello,

I used to use Prisma alot with Apollo, and I'd like to completely replace Prisma with Cruddl, but I have a few questions:

• Is there a typeDefs generation for Cruddl as Prisma does (https://www.prisma.io/docs/prisma-cli-and-configuration/cli-command-reference/prisma-generate-xcv2/)

The reason for asking this is because currently (as far as the docs go), I'd have to setup a GraphQL

Trying to achieve multitenancy using the documented SpEl expression, but the expressions seem to be evaluated before Spring Security filters.

Is there any possibility to use Authentication / Principal in @Document("XXX_MyDoc") expression?

one of the cluster tests is to start a 4th instance that is suspected to give up and exit once it failed to start the arangod a hundret times. This doesn't happen, process continues to run.

However, the solisoft.gitbooks.io domain version seems to work correctly.

In foxy.ovh, when you click to read a chapter, this message is displayed, and then the UI becomes unresponsive and the user can't navigate around.

velocypack needs to be installed on the target system it should be added to the repository or mentioned in the install instructions.

Implement more methods of the ArangoDB's REST API marked for milestone 2 (M.2).

The rincon_client crate does not implement all methods of ArangoDB's REST API. The current status is documented in docs/arangodb_rest_api_methods.md. Methods that are planned to be implemented are marked with a milestone.

The way we handle authorization in our resolvers is pretty nice and declarative. We should open source this module and perhaps even write a little blog post.

https://github.com/mikestaub/slack-lunch-club/blob/master/backend/packages/graphql/schema/utils.js