Santa wish list for Dgraph in 2021

[mrjn (Manish R Jain)]

Let’s use this thread to determine what features we’d like to see in Dgraph in 2021. I’ll start:

• Multi Tenancy
• Change Data Capture
• Integration with Kafka
• Ability to load data directly from SQL to Dgraph Cloud offerings
• Query Planner
• Read-only replicas
• Scale to 100 TBs of data

Anything else? Let us know!

[amaster507 (Anthony Master)]

• Boolean support in RBAC
• Filtering parent Nodes with Children properties without cascade
• Array support (duplicate and orderable)
• JS hooks pre/post generated queries/mutations. (Not just replacing with lambda resolvers)
• Enhancements on ordering (not sure of best solution here yet)
• String functions
• Datetime functions
• Pagination with Cascade support

[iluminae (Kenan Kessler)]

1. Multitenancy (we are doing this ourselves with a prefix and it’s not super) Doing this in the foss version would be the real gift.
2. gRPC streaming mutations and gRPC streaming responses. Heck just gRPC the whole shebang.
3. string manipulation functions would be a nice to have.
4. we have to do a sad amount of extra work to filter edges by lists of (time)ranges. In the current system that would mean supporting lists as facets and supporting a range predicate with filters that could work on it… Ok that’s a big one but it would remove 2/3 of our graph.

[aman-bansal]

@amaster507

Boolean support in RBAC

@amaster507 Christmas came early for you . This has already been merged in master and should go out in next release

Also I have seen this requests that will be great to see:

1. Distributed predicates
2. Continue bulk/live from where is stopped earlier
3. Composite indexes like unique on composite fields

[MichelDiz]

My only wish in 2021 is Loops in DQL!

[trka (Kevin Ard)]

I’m a noob, so mine could all be there and I just haven’t noticed yet lol.

• @recurse on non-root blocks would be tops
• hot re-index. Even at some expense, but not stop-the-world.

[mrjn (Manish R Jain)]

Re-indexing already allows a “background” option. Correct me if I’m wrong, @ibrahim

[chewxy]

default values support for graphql. Right now you can do it with lambdas but it’s a pretty crazy thing to do.

[Willem520 (Yu Wei)]

Multi Tenancy
CDC
Integration with Kafka
these are more important for me. and the system stability is also important.

[trka (Kevin Ard)]

Ahhh… Could be a docs issue. I’d have to find the part that led me to that (or… That I misread). Trying to find it again, I see some changes in 1.12, and MR#4819 with bg index.

My Santa came early. …last spring lol.

[verneleem]

default values with a directive would be much better. Right now a default value with lambda requires a custom mutation which then you lose the benefit of creating deeply nested data in a single mutation. You couldn’t for instance create a author with a post and have a default value on the post unless now you created two lambda mutations, one for just a new post and one for a new post with author.

[wildan2711 (Wildan Maulana Syahidillah)]

• [DQL] Sort indexes, or anything to improve the below issue:
  Offset-based pagination is slow - #6 by diggy
  Usecase: user feeds with infinite scroll
• [DQL] Random nodes function
  Pick random node - #9 by EnricoMi
  Usecase: friend recommendations
• [DQL] tf-idf scoring on term search
  Feature request: full text search with tf-idf Scoring - #8 by diggy
• [DQL] Unique indexes
• [DQL] Composite indexes
• [DQL] Geo Distance function
  Add @distance for geo

[ibrahim (Ibrahim Jarif)]

@mrjn

Re-indexing already allows a “background” option. Correct me if I’m wrong, @ibrahim

That’s correct.

@trka https://dgraph.io/docs/query-language/schema/#indexes-in-background is what you’re looking for

[zzl221000 (JimZhang)]

• Facets variable in Upsert Block
  The edges cannot be copied correctly now.

[amaster507 (Anthony Master)]

Deep Patch mutations.

References:

Cannot update multiple nodes at once

context: Updating multiple nodes at once with their own separate patches. I can generate a mutation with multiple mutation statements one for each node, but my IDE frowns upon dynamic queries/mutations in this regards. I can’t do this with the generated updateType mutation as it stands because it only takes one filter and patches all nodes that match the filter with the same patch. I thought I could do this using a deep mutation, but deep mutations do not allow you to patch child nodes. type…

Deep Mutations - Graphql

Mutations also allows to perform deep mutation at multiple levels. We use the following schema to demonstrate some examples. Schema: type Author { id: ID! name: String! @search(by: [hash]) dob: DateTime posts: [Post] } type Post { postID: ID! title: String! @search(by: [term, fulltext]) text: String @search(by: [fulltext, term]) datePublished: DateTime } Example: Deep Deep mutation using variables mutation DeepAuthor($author: DeepAuthorInput!) { DeepAuthor(input: [$author]) { …

[billybobthorton (Billy)]

Ability to do atomic mutations without DQL

[Valdanito]

GraphQL has too many restrictions. I hope DQL can be more powerful and compete with the top graph database.

[musiciansLyf]

1. I hope there can be synchronous transactions, which can directly sense that the transaction has been completed.

2. More detailed log information, such as downtime log.

3. Java connection management: after the node is down, the connection of the node will be skipped. At the same time, the connection will be restored after the node is restored.

4. The cluster can have higher stability for concurrent requests. Now there will be node downtime and connection exception.

[iluminae (Kenan Kessler)]

One more for me: Ratel… needs love.

[andrewluetgers]

I would love to see in Slash GraphQL

• Support for more AWS regions, at least add in US East
• AWS VPC Peering, the Mongodb Atlas service does this very well