3 ms·
A similar philosophy has made the "tidyverse" a much-loved extension of the statistical language R. Compare the following 2 equivalent snippets. Which one seem
by antipaul 6y ago
A similar philosophy has made the "tidyverse" a much-loved extension of the statistical language R.
Compare the following 2 equivalent snippets. Which one seems more understandable?
iris_data %>%
names() %>%
tolower() %>%
gsub(".", "_", ., fixed=TRUE) %>%
paste0("(", ., ")")
or:
paste0("(", gsub(".", "_", tolower(names(iris_data)), fixed=TRUE), ")")
- chubot 6y agoYup, I'm working on Oil shell and I wrote an article about data frames, comparing tidyverse, Pandas, and raw Python: What Is a Data Frame? (In Python, R, and SQL) http://www.oilshell.org/blog/2018/11/30.html http://www.oilshell.org/blog/2018/11/30.html The idea is essentially to combine tidyverse and shell -- i.e. put structured data in pipes. I use both shell and R for data cleaning. My approach is different than say PowerShell because data still gets serialized; it's just more strictly defined and easily parseable. It's more like JSON than in-memory objects. The left-to-right syntax is nicer and more composable IMO, and many functional languages are growing this feature (Scala, maybe Haskell?). Although I think Unix pipes serve a distinct use case.