You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
not sure exactly what features you would want in it so i kept it simple . we can add index on the key for a more advanced store, but honestly if it is only key->value lookup that is being done, the K field should be removed. I copied the mongodb test suite and it seems to pass, so that's good at least.
arangodb is a bit weird in its rules for byte arrays and keys. this is explained at the top of arangodb.go. tl;dr keys need to be a subset of ascii, and there is no 1 byte numeric type (only 8).
as for connecting, it creates the database + collection on connection if they do not exist, and so it would be up to user to either use a user with permissions to do so, or to pre-populate their database.
let me know if there is anything else that would be needed, or if you have any questions.
arangodb is just a nosql database layer on top of rocksdb, so performance wise, raw rocksdb will of course always be better.
Hello 👋 , Wow thank you for working on this! I had a quick look just now and it looks pretty good 👍 . You can open a PR, then I will dig into the details and maybe ask some questions around individual lines.
ArangoDB seems popular due to its multi-model approach, although key-value pairs are currently only implemented as documents with single values and no secondary indexes (see arangodb/arangodb#663 and https://docs.arangodb.com/3.4/Manual/Architecture/DeploymentModes/Cluster/DataModels.html#keyvalue-pairs).
The text was updated successfully, but these errors were encountered: