I'd love to use data.world as a backend for SQL but have it work in R Notebooks via the knitr {sql} engine. That enables students to simply write SQL and not have to worry about the R code around it.
Two ways forward I think:
A datadotworld knitr engine (replacing the sql engine). I tried that but run into some trouble because R Notebooks via RStudio worked differently than straight Knitr. I posted about that on a comment at: http://rmarkdown.rstudio.com/authoring_knitr_engines.html (might still be in moderation?)
Implement a (limited) DBI backend for the data.world API. That would be broadly useful but in this specific case it would make the knitr sql engine "just work". The basics of constructing a DBI compliant backend look like it might be possible?
https://cran.r-project.org/web/packages/DBI/vignettes/backend.html
I'd love to use data.world as a backend for SQL but have it work in R Notebooks via the knitr {sql} engine. That enables students to simply write SQL and not have to worry about the R code around it.
Two ways forward I think:
A datadotworld knitr engine (replacing the sql engine). I tried that but run into some trouble because R Notebooks via RStudio worked differently than straight Knitr. I posted about that on a comment at: http://rmarkdown.rstudio.com/authoring_knitr_engines.html (might still be in moderation?)
Implement a (limited) DBI backend for the data.world API. That would be broadly useful but in this specific case it would make the knitr sql engine "just work". The basics of constructing a DBI compliant backend look like it might be possible? https://cran.r-project.org/web/packages/DBI/vignettes/backend.html