Hacker News Reader

Comment by ggm | Hacker News Reader

I tried R with JSON and R with Elastic Search and both interfaces were 'abandonware' with 'just use the HTTP paradigm and deal with the data frame afterward' outcomes.

Felt like a really low bar, to get real-world data forms imported, and find the <- deplyr functional bindings pretty much 'not there'

I liked Shiny, because of the low barrier to deployment of a GUI inside the company leveraging R for the graphing. But integration is .. painful.

View on HN

Were you using jsonlite? Great library that can parse directly from a URL and transform into data.frame depending on JSON representation. I agree it is not clear from the get go which library should be used when you do something outside of base R. You kind of need to be in the known

I think I was, yes. This is 2+ years ago. I haven't gone back because I got there with REST fetches and internalized re-framing of the JSON. I was just amazed how 'crude' it was, and that the dplyer bindings for ElasticSearch were non-functional (for me at least)

The elastic library on CRAN got a major update a few months ago which made it much more powerful and easier to use. Makes uploading easy too, especially if there's a direct relationship between the structure of your dataframe and the Elastic json. Uploading nested structures is still a bit of a pain though.

Thanks. Good to know, we may re-visit things. Elastic itself is a pain to manage, and with data growth we're re-examining what we get from it. For time series data, there is an element of "why?" about it: PostGres as a time series JSON blob store, might work better for us. The sharding by time is easy to understand and it has more affinity to the underlying host OS. Elastic, you have to understand a LOT of moving parts to make it work well (or be a Kubernetes expert, or pay somebody else) -if you already have an investment in PG then its possibly better use of your time.

But its good to know the interface got some eyeballs. I retract the accusation of abandonware.

Reply on news.ycombinator.com