R vs. Python for Data Science: Summary of Modern Advances

elitedatascience.com

27 pointslizeds4 comments

Both R and Python have gained libraries for SQL syntax queries on their datasets. These are powered by sqllite. As much of the data that gets into R or Python is obtained from a SQL query (or SparkQL etc.) it's awesome to be able to query, aggregate and even join internal datasets in the same way! https://github.com/ggrothendieck/sqldf is game changing.

The wealth of statistical packages available for R is astounding. Packages like GLMNet are made available by the top researchers in the field. In comparison, Python libraries tend to be relatively unsophisticated and less robust.

Sqldf is very tempting, but at least in the past it was so much, much slower than dplyr that the time to learn dplyr's syntax could pay off in a single project, or possibly even a single large operation.

Agreed. The ability to query with SQL syntax has been great.

It covers many of the libraries discussed here:

https://news.ycombinator.com/item?id=13110230

Reply on news.ycombinator.com