pubkey/rxdb

Torsten Raudssus

@Getty

the shortage of my old company name was RSS

well i thought at first its kinda genius company name cause of the good SEO rating ;-)

pubkey in a domain name would probably instantly go onto the first page for that term

pubkey

@pubkey

@/all Help me to improve RxDB by answering a few questions ( < 1 minute):
Open the survey under https://forms.gle/y9d73Npyu8h84Ep29

Joseph Shearer

@jshearer

@pubkey Question for you... Before I go and create a fully reproducible test case, do you have any idea why querying for rows that match a selector (.find({key: "value"}).exec()) will return the correct row only when that field is not indexed? As soon as I indicate it as index: true, I get no results for my query. I tried dropping down into collection.pouch.createIndex and collection.pouch.find, with the same results: adding an index for a field prevents me from querying for that field. If I get all documents, my document is there as expected, with the correct field value.

The same is true for sorting (which is where we initially encountered this issue) -- adding an index for a specific field means that sorting a query by that field yields no results

Joseph Shearer

@jshearer

It looks like I encountered the same exact scenario as @rmisio described if I search for this issue in the logs here: sorting doesn't work without indexes, add indexes, then sorting by field returns no results, as does filtering by that field.

pubkey

@pubkey

@jshearer this might be related to pouchdb/pouchdb#7810

which adapter and version do you use?

pubkey

@pubkey

I think you should create an issue with code to reproduce. So we can create a test which pins the correct behavior down

pubkey

@pubkey

@/all Released version 8.4.0 will two bugfixes: https://twitter.com/rxdbjs/status/1168265902720278528

100 Errors

@theuser_gitlab

is rxdb stable?

pubkey

@pubkey

@theuser_gitlab depends on how you define it

but yes

people use it in projects

Torsten Raudssus

@Getty

I still don't know how to start with it ;)

pubkey

@pubkey

https://github.com/pubkey/rxdb/tree/master/examples

100 Errors

@theuser_gitlab

can anybody create databases or are there any limitations?

Torsten Raudssus

@Getty

@pubkey i still dont know where to start ;-) My problem is primary that i am coming from a mixed environment, so i got an actual non-JS server which should be involved. Somehow I thought i can write a node-server with RxDB and just access it from my JS and from my non-JS server code, but now i am looking at the vanilla example, and i have a hard time understanding what exactly i am looking at given that its all just still one codeblock.

I also wanna be clear here that i dont expect the non-JS server to be able to do the full RxDB featureset, but at least i should be able to make in this node server i talked about the code to take in some HTTP to update and change data. At least thats the thought

what especially confuses me is that in the vanilla example the createDatabase is in that same JS code that updates the HTML, that is.... confusing

Torsten Raudssus

@Getty

actually i could add up a V8 on the non-JS server, that could make the situation easier

pubkey

@pubkey

@Getty I am not clear about what problem you try to solve by using RxDB

When your sever is non-JS, I recommend to create a GraphQL endpoint with which you can sync the clients. also you could write to a couchdb from whatever backend and use that to sync

Torsten Raudssus

@Getty

well my thought is that RxDB can provide the central point where on one side the app is just updating the state and on the other side the clients get their updates with the live update. If I would have a GraphQL subscription, then I would go totally around RxDB, or? I might be just confused about the position of RxDB in the world 8-D

pubkey

@pubkey

You can replicate the data from a graphql-endpoint into the clients RxDB. Then your queries are faster because you do not have to go throught the internet and also your app can be used offline

Torsten Raudssus

@Getty

OOOHHH that might be my central misunderstanding

RxDB is just a client thing?

100 Errors

@theuser_gitlab

as far as I understood not

on client side you can use pouchdb

( https://pouchdb.com )

100 Errors

@theuser_gitlab

do I need to create a custom build for authentication etc? https://rxdb.info/custom-build.html

pubkey

@pubkey

you can use RxDB on every javascript runtime. this might be the browser or the nodejs server.
@theuser_gitlab custom builds have nothing to do with authenticatino

100 Errors

@theuser_gitlab

I mean that I build my custom authentication with a custom build

or is there another way to do this?

Unforgiven-wanda

@Unforgiven-wanda

Hello,
Question about feasibility: Could I use an RxDB collection to power a chat room, by replicating the messages between the participants? Would you know if anything similar has ever been attempted?
Thank you.

pubkey

@pubkey

@Unforgiven-wanda yes that can be done with rxDB. In face chatapps are the most used demo-showcase for realtime technologies. I also have done that, for a evaluation I implemented the same chat-app (similar to whatsapp.web) with different realtime technologies.

100 Errors

@theuser_gitlab

hi, is there something like https://pouchdb.com/api.html#fetch_document options.revs_info or options.open_revs available in rxdb?

pubkey

@pubkey

@theuser_gitlab not via RxDB but you can always use the internal pouchdb like myRxCollection.pouch

100 Errors

@theuser_gitlab

ok thanks

Unforgiven-wanda

@Unforgiven-wanda

@pubkey Got it. Thank you.

David Welling

@DaveWelling

Hello, I'm new to rxdb, and I was curious if the primary motivation for schemas is data-integrity or does it also facilitate other features of the database (such as change observation)?

pubkey

@pubkey

@DaveWelling
Here are some other reasons:

There are possible optimisations than can be used when a query is observed. Some of them only work with a schema.
The key-compression is only possible with a schema.
For the encryption-feature it must be known which fields are encrypted and which are not.
Defining indexes is easier when there is a schema
For the data-migration feature it is easier to have a schema which shows when things have changed
the population-feature only works with a schema

The schema-validation itself is optional and mostly used in dev-mode because of the performance and increased build-size of the validator

David Welling

@DaveWelling

That's a super-helpful answer. Thank you very much!

Unforgiven-wanda

@Unforgiven-wanda

Hello again
Assuming I have two or more users, all replicating the same collection - let's call it notes- how would you handle the conflicts that are bound to happen when some of them go offline, edit their documents, then come back online again to sync them? Should we handle this server-side or client-side and how? I came across https://github.com/jo/pouch-resolve-conflicts but I'm not too clear on how I can use it with RxDB
Any pointers or code example would be most welcome.
Thank you.

pubkey

@pubkey

@Unforgiven-wanda there is no conflict resolution in RxDB, rafamel/rxdb-utils#5
You can do run conflict resolution by subscribing to changes and on every change get the document from pouchdb with conflicts: true

Unforgiven-wanda

@Unforgiven-wanda

You wouldn't happen to have an example on hand, would you? ^^

pubkey

@pubkey

no sry

Where communities thrive

People

Repo info

Activity