Re:Using PostgreSQL’s JSONB for NoSQL (2019) [video]
by stevesimmons on Sunday September 05, 2021 @05:35AM
(#28422790)
Does anyone know of performance (and usability) comparisons between say Postgres JSONB and Azure CosmosDB?
[ Parent
| on HN ]
Re:Using PostgreSQL’s JSONB for NoSQL (2019) [video]
by stevesimmons on Sunday September 05, 2021 @07:57AM
(#28423623)
To expand on my particular use case for Azure CosmosDB, I have a Python web app built with FastAPI. I find it super convenient to dump Pydantic data models of our primary domain objects to JSON and save them into Cosmos, tagged with the model name.
Then on rereading the Cosmos records, my model registry automatically imports the required Pydantic model and reconstitutes the live Python object.
Re:Using PostgreSQL’s JSONB for NoSQL (2019) [video]
by fisf on Sunday September 05, 2021 @07:29AM
(#28423445)
We use azure SQL (basically mssql) to perform something similar -- this might be a more apt comparison.
We deal with largish amounts of JSON event data, and unpack specific fields to computed columns. This allows to use clustered indices, standard analytical tools, etc., while leaving the raw data as it is.
Performance is mostly fine for analytics (time series aggregation and such). It is also a very flexible approach.
Re:Using PostgreSQL’s JSONB for NoSQL (2019) [video]
by jf22 on Sunday September 05, 2021 @06:53AM
(#28423228)
Those are two completely different tools with different use cases.
Re:Using PostgreSQL’s JSONB for NoSQL (2019) [video]
by exceptione on Sunday September 05, 2021 @06:57AM
(#28423256)
Well, tell that to the MS-shops... I think it is a very good question, would love to see some benchmarks.