Accessing components of data from the "." (dot) operator with magrittr

Question

I'm not clear on how to apply functions to components of data using the dot (".") with magrittr such as columns of a data from or items in a list.

Example:

> data.frame(x = 1:10, y = 11:20) %>% .$y
[1] 11 12 13 14 15 16 17 18 19 20

It seems that accessing the data should work the same as applying a function to it, but it does not:

> data.frame(x = 1:10, y = 11:20) %>% min(.$y)
[1] 1

MrFlick · Accepted Answer · 2017-07-19T20:54:31.243

3

The data.frame will be passed as the first parameter unless a lone dot is placed somewhere else in the call.

data.frame(x = 1:10, y = 11:20) %>% min(.$y)

is the same as

dd <- data.frame(x = 1:10, y = 11:20)
min(dd, dd$y)
# [1] 1

You would have to use a code block

data.frame(x = 1:10, y = 11:20) %>% {min(.$y)}

edited Jul 19 '17 at 20:54

answered Jul 19 '17 at 18:52

MrFlick

Thanks. I'm curious, then, why data.frame(x = 1:10, y = 11:20) %>% .$y doesn't give something like data.frame(dd, dd$y). Also, doesn't my first example contain a "lone dot placed somewhere else in the call"? – jzadra Jul 19 '17 at 20:52
Because `.$y` is equivalent to `'$'(., y)` which is already putting the data.frame where you want it. You could also use `%>% '$'(y)` without the dot. It's really more about not going into nested calls than a "lone dot" so I was being a bit imprecise there. – MrFlick Jul 19 '17 at 20:58
1

@jzadra Btw, there's some esoterics associated with how magrittr behaves when `.` appears explicitly as an argument to the function. Namely, if it is a freestanding argument to the function (as MrFlick showed is the case in your example), then the dot is not additionally given as the first argument. So, for example `"A" %>% setNames(.)` is not equivalent to `setNames("A", "A")` and in fact fails. https://github.com/tidyverse/magrittr/issues/128 (Oh, just noticed MrFlick edited the comment to cover that. Anywho, I'll leave this comment here.) – Frank Jul 19 '17 at 21:04
For this exampe, you may want to use : `data.frame(x = 1:10, y = 11:20) %$% min(y)` – moodymudskipper Aug 15 '17 at 15:51

1 Answers1