You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For a while now I've been rather insistent that the top level of a dataset be an array. I've now come across a number of examples in the wild where having an object as the top level is just plain necessary. For example:
data.json catalogs
top level is an object. It's a dataset. qri should get over it.
site maps
Doing de-duplicated sitemaps is much easier using normalized urls as keys in an object.
So based on this, I think we need to relax the definition to say that a top level dataset needs to be either an array or an object. This'll affect a few things:
dsio.Reader will need to recognize object keys as entries
any validation needs to be cool with { "type": "array" } or {"type":"object"} as a root schema
I'll add to that list as things go.
The text was updated successfully, but these errors were encountered:
For a while now I've been rather insistent that the top level of a dataset be an array. I've now come across a number of examples in the wild where having an object as the top level is just plain necessary. For example:
data.json
catalogstop level is an object. It's a dataset. qri should get over it.
site maps
Doing de-duplicated sitemaps is much easier using normalized urls as keys in an object.
So based on this, I think we need to relax the definition to say that a top level dataset needs to be either an array or an object. This'll affect a few things:
{ "type": "array" }
or{"type":"object"}
as a root schemaI'll add to that list as things go.
The text was updated successfully, but these errors were encountered: