Exists

[vulcandragi]

Add a variant of find, which checks if the record exists in the database but doesn't return any data, just says whether or not it exists

This improves performance in some cases where a test of whether the user has already been registered and the case will not be registered

There are better ways to do this test, which is in the insertion case of failure, check the error code, but as it is not so easy at the moment to obtain this code, it would be easier to implement this function

[RomainMazB]

This would be useful indeed, I may take a look to implement this in the next few days.

[billy1624]

How about adding a new operation exists under find?

let _: bool = Entity::find_by_id(1)
    .exists(db)
    .await?

let _: bool = Entity::find()
    .filter( ... )
    .exists(db)
    .await?

[cemremengu]

Was this ever implemented or still up for a PR?

[vulcandragi]

I think it's good like this, just one thing that maybe you should take into account, when using the filter maybe there can be more than one, but when the filter is by id, it can only exist or not

So maybe you can use count(1) in the filter to be able to nab the amount of existing records

[billy1624]

Btw... you could also count it with paginator
https://docs.rs/sea-orm/0.4.2/sea_orm/trait.PaginatorTrait.html#method.count

[vulcandragi]

yes, I agree with what you said, but that would be one more call to the database

that's why I talked about the filter returning the number of records, but it's just my opinion

[RomainMazB]

🤔 Using the sql EXISTS statement, no matter if there is one or multiple results, it will just return TRUE or FALSE directly from the database.

I thought this would be already implemented into sea-query but it's not. I guess this should be the starting point.

Related to SeaQL/sea-query#118

[vulcandragi]

makes sense, and when I did that I didn't use this command, but looking at it from this side it only makes sense to return the boolean value

[tyt2y3]

I am still thinking whether we should add the method to PaginatorTrait or create a new Trait (which includes other methods like max or min)

[vulcandragi]

I think it would be interesting to be separated as it doesn't have anything to do strictly with pagination, but something useful that can be used in any context.

[penso]

I'm currently slowly switch a Ruby/ActiveRecord API to Rust and would love to see an exists(db) returning true/false. The Ruby AR query looks like this: SELECT 1 AS one FROM your_model LIMIT 1 which could be the most effective.

[ikrivosheev]

Related to: SeaQL/sea-query#118

[Xiaobaishushu25]

Can the exists function be used in seaorm now? Do you have any simple examples?

[tyt2y3]

You can use the async fn count() from PaginatorTrait.

[IgnisDa]

@billy1624 @tyt2y3 Now that it has been merged to sea-query, can we expect an .exists() function? .count() is not as performant as exists.

[Expurple]

I can't say on their behalf, but I'm a new co-mainainer and we're currently working on the next major release of SeaORM and SeaQuery. So, I prioritize breaking-change ideas and PRs higher. I assume that this will go on for a couple more months.

I have definitely wished for .exists() many times in my own codebase, and this feature is on my own long-term todo list.

TL;DR we'd like to have this, but without outside contributions it probably won't happen earlier than months from now.

[IgnisDa]

If you could tell me where you'd like it implemented, I'd love to make a PR for it.

…

On Thu, Jun 12, 2025, 11:55 PM Dmitrii Aleksandrov ***@***.***> wrote: I can't say on their behalf, but I'm a new co-mainainer and we're currently working on the next major release of SeaORM and SeaQuery. So, I prioritize breaking-change ideas and PRs higher. I assume that this will go on for a couple more months. I have definitely wished for .exists() many times in my own codebase, and this feature is on my own long-term todo list. TL;DR we'd like to have this, but without outside contributions it probably won't happen earlier than a few months from now. — Reply to this email directly, view it on GitHub <#730 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AOQ5PNF5SS4F4RW4XBE234T3DHA3DAVCNFSM6AAAAAB7EERCUOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNBVGIYTANY> . You are receiving this because you commented.Message ID: ***@***.***>

[Expurple]

I'm not familiar with this area myself. But I'd start with looking at the similar async SeaORM methods that consume the statement and execute it in the provided connection/transaction. Methods like count, one, all.

count is in PaginatorTrait.

For some reason, one and all are duplicated inherent methods in Selector, SelectorRaw, and Select* types. This looks like a design flaw and something that should be abstracted into a trait. Added this to #2548

[IgnisDa]

@Expurple Made an initial PR in #2623. Would appreciate your feedback on it.