Log file Viewer

[00:23:52] <GothAlice> Nope. Use relational structures when you have highly relational data, optionally with complex constraint requirements. Use EAV if you have highly irregular data, i.e. a use case people are often tempted to try to apply to document stores, badly. Use document stores if you have rich, nested structures to model, or data needs with flexible (not irregular) requirements.

[00:24:14] <GothAlice> jl-: I like to point people at https://www.javaworld.com/article/2088406/how-to-screw-up-your-mongodb-schema-design.html as a nice introductory summary.

[00:34:50] <jl-> GothAlice: thanks I'll give that a read

[00:52:19] <jl-> GothAlice: what is EAV?

[01:58:41] <GothAlice> jl-: https://en.wikipedia.org/wiki/Entity–attribute–value_model

[01:58:50] <GothAlice> Think of medical data, like lab results.

[02:00:20] <GothAlice> There are a bajillion different proteins, and enzymes, and hormones, and pH levels, and … that you need to boil the association of data values to entities ("records", i.e. "the lab results of requisition X") down to the absolute bare minimum, as no two reports will be the same.

[02:01:21] <GothAlice> (Classic "sparse attributes" problem.)

[09:22:30] <mainnet> Hey! I'm just here to find out if we can we add two rs members to an already existing replica set (3 member) without downtime?

[09:23:02] <kali> mainnet: yes :)

[09:23:55] <kali> mainnet: basically just do it. i recommend you let one resync completely before adding the second one not to stress you cluste too much

[10:27:17] <mainnet> kali, just...just do it? As simple as that?

[10:28:53] <GothAlice> mainnet: When in doubt, try it out, and "do you have backups?" means "I can't fix this", not "this can't be fixed". ;^P

[10:41:01] <mainnet> GothAlice, I do. :D I'll give it a try.

[23:59:23] <cslcm> Hey. Any idea why i'd get a "no documents in result" result from a FindOneAndUpdate action with Upsert?