99 results found
-
Vector embedding search should include distance in results
The newly released capability is https://firebase.google.com/docs/firestore/vector-search is working very nicely! However, all similar products should return the distance metric in the query results (which the backend has no doubt already calculated) so that the client perform further analysis based on that distance
7 votes -
Allow listening to firestore document/query changes without getting a snapshot automatially
The default behavior when listening to a document/query is that you will always get an initial invoke of your listener.
Allowing to change this behavior would give more freedom, in the case that the developer wishes to fetch the initial values asynchronously and then have the changes delivered via the callback.2 votes -
Support offline capabilities for mobile using React Native
I can see offline capabilities for mobile using Kotlin, Java, Swift and Objective-C. I am requesting offline support for React Native application.
1 vote -
Raise the Firestore Rules Expression Limit
Raise the maximum expression limit for firestore rules to make more complex doc schema validation possible.
1 vote -
Transactions should allow exponential backoff retries like pub sub
Transactions allow passing an options object with maxRetries, but all retries are immediate. If there are many other functions writing the same document, transactions will fail after all retries (lock will time out). Maybe a better strategy would be to delay with exponential backoff like pub sub does.
3 votes -
custom gpt to firestore
I wanted to create my own GPT using OpenAI's new GPT builder which has "Functions" in it, and I wanted it to be able to use Firestore API to perform CRUD operations, but unfortunately I struggled to find documentation on how to do that.
2 votes -
Allow developers to use the local Firestore database for free
It'd be really nice if you could allow users to create local drafts, of small documents, for instance, and save them locally for free with Firestore.
This way you wouldn't have to coordinate Firestore and its rules with some other database and client code nor would you have to choose the simpler option of getting charged for writes and reads on remote database documents that really only one user is supposed to write and read.
Please note that while you could use the realtime database to save drafts, 1 it's not always easy to figure out whether that'd actually be…
1 vote -
Support offset-based document access and pagination
There should be an easy way to get a specific document by index in a query:
// this query should work as it's read, if i wanted to use gte i'd use a where
query(this.collection, orderBy('anything', 'asc'), startAt(index), limit(1))
3 votes -
Firestore : Throw error access data with offline cache without enable persistence
When retrieving documents from Firebase's Firestore on the web, if you attempt to read the value from the cache, it will return an empty result.
This seems to be an issue because Firestore's database does not use offline data when accessed from the web by default.
To make developers aware of this, should throw some error if Firestore's database is set to not use offline data, and attempting to read the cache value.
3 votes -
Improve firestore explorer UI - it is too small
The UX when reading large objects in the firestore explorer UI requires too much scrolling in small windows to find the information you are after. I'd propose to:
Note: "window" here refers to where the container where the data of a document can be found or alternatively where all the collection can be found. So the firestore explorer contains usually, a collection window, a document window, (a sub collections window, a document window,...).
- collapse maps by default, or collapse maps if the object is larger than the current window. When dealing with large objects it's currently annoying to find the…
6 votes -
Duplicate document in collection in Firestore
Duplicate document in collection in Firestore
3 votes -
Firestore return document when it gets updated
As we get the whole document when it is created, it would be nice to get it as well when it gets updated.
Example:
Online game where 4 people join together. In document, we have a list of players with their names. These names should be displayed for all players. Last user joins closing the session. To get the rest of the names after update the document with this last player, I need to do an extra read.Problem:
I am making an entra call for every player, affecting my quota limitation.Solution:
Return updated document.9 votes -
Add Creator User UID to the Collection Record Metadata
Will allow further security validation and early detection of security checks failures.
To give an example of why this is important,
I use a Firebase Collection as an API alternative, I create a task record that triggers a cloud function, it allows me to manage my resources effectively and provide partial updates, way more flexible than traditional REST, however with API I use bearer tokens to verify that the identity of the person creating the record and deciding if they are allowed or 401, without UID in record metadata I have to send the plain UID as a field in…2 votes -
Firestore hosting in South Africa
Make Firestore and Firebase hosting available on the new Johannesburg Cloud Region; this will greatly assist developers in launching apps within the continent and also assist with local data protection regulations.
2 votes -
Ability to disable deleting collections and documents from firebase console
In Firebase's Firestore, a user with write access to a project can simply delete any document and even a whole collection with multiple documents with a single click. This is a major security problem as any of my team member may end up deleting a whole collection with millions of documents in under a minute, through the Firebase console.
I would like to request for a feature to disable this.
To be clear - The possibility to grant write access without delete access. Currently, if I am not mistaken, through the User and Permissions panel it's kind of "all or…
10 votes -
Support deleting databases via the API
You cannot delete a Firestore database with the API; you must use the CLI if you want to do this programmatically. This is problematic as it means you essentially need to create a shell in your application, ensure the CLI is available, authenticate, and then make the required CLI call.
2 votes -
Support deleting collections in the API
Currently you can only delete collections programmatically via the CLI. Otherwise, you have to list and delete every single document in the collection. If those documents have nested collections, you have to (recursively) retrieve and delete every document in the nested collection (As deleting a document does not delete any nested collections the document has). If there are additional levels of nested collections... You get the idea.
I wasn't sure what this other request was referring to so I made this one: https://firebase.uservoice.com/forums/948424-general/suggestions/46562317-api-support-for-deleting-directories
7 votes -
Expose subcollections as json fields
Currently, subcollections are treated separately from fields in the UI. This seems like an arbitrary distinction - a subcollection is simply a json object under a document. You could do the same thing with a string field.
The API could support retrieving subcollections as string fields in json format for a document. This would prevent the need to iterate over every document in a collection, then get a list of the subcollections, then retrieve the documents from those subcollections, etc. If worried about time required, since Firestore supports 100 nesting levels, perhaps the API call could include the nesting level…
1 vote -
Expose "last updated" metadata
If Firebase cannot support some kind of changestream implementation, it should include metadata on each document that includes a "last updated" timestamp. Then clients can query against this field to only retrieve "new" changes (Inserts, Updates). It lacks support for delete capture but is the next best thing after a changestream.
18 votes -
Changestream
Firestore does not support any actual CDC (Change Data Capture). A changestream similar to other NoSQL implementations (Mongo, DocumentDB, Azure Cosmos DB) would make the database a lot more usable and grant feature parity with other popular NoSQL databases. This changestream would expose operations like updates, inserts, and deletes for a client to come and retrieve from either a timestamp or a resume token.
7 votes
- Don't see your idea?