A wee look at group_map and group_split in dplyr

[This article was first published on HighlandR, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Dplyr 0.8.0 launched recently, which you probably already know, but just in case you missed it..

Two new functions have been catching my eye : group_map and group_split.

The aim of this post – take a first look at these and try and get a new blog post up on github before February is out.

Time now: 23:38(UK)

What it says on the tin

Load packages and create a function for demo purposes , basically grab the first 5 rows of whatever we throw at it:

library(nycflights13)

## Warning: package 'nycflights13' was built under R version 3.5.2

library(dplyr)

## Warning: package 'dplyr' was built under R version 3.5.2

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

data(flights)

header <- function(x,n = 5){
  head(x,n)
  }

Let’s check the function works first of all:

## # A tibble: 5 x 19
## # Groups:   origin [3]
##    year month   day dep_time sched_dep_time dep_delay arr_time
##   <int> <int> <int>    <int>          <int>     <dbl>    <int>
## 1  2013     1     1      517            515         2      830
## 2  2013     1     1      533            529         4      850
## 3  2013     1     1      542            540         2      923
## 4  2013     1     1      544            545        -1     1004
## 5  2013     1     1      554            600        -6      812
## # ... with 12 more variables: sched_arr_time <int>, arr_delay <dbl>,
## #   carrier <chr>, flight <int>, tailnum <chr>, origin <chr>, dest <chr>,
## #   air_time <dbl>, distance <dbl>, hour <dbl>, minute <dbl>,
## #   time_hour <dttm>

Good, that just returns 5 rows from the entire datset. Now let’s replace group_by with group_split, and group by origin

split_test <- flights %>% 
group_by(origin) 

Then we use group_split to create individual tibbles for each group

group_split(split_test)

## [[1]]
## # A tibble: 120,835 x 19
##     year month   day dep_time sched_dep_time dep_delay arr_time
##    <int> <int> <int>    <int>          <int>     <dbl>    <int>
##  1  2013     1     1      517            515         2      830
##  2  2013     1     1      554            558        -4      740
##  3  2013     1     1      555            600        -5      913
##  4  2013     1     1      558            600        -2      923
##  5  2013     1     1      559            600        -1      854
##  6  2013     1     1      601            600         1      844
##  7  2013     1     1      606            610        -4      858
##  8  2013     1     1      607            607         0      858
##  9  2013     1     1      608            600         8      807
## 10  2013     1     1      615            615         0      833
## # ... with 120,825 more rows, and 12 more variables: sched_arr_time <int>,
## #   arr_delay <dbl>, carrier <chr>, flight <int>, tailnum <chr>,
## #   origin <chr>, dest <chr>, air_time <dbl>, distance <dbl>, hour <dbl>,
## #   minute <dbl>, time_hour <dttm>
## 
## [[2]]
## # A tibble: 111,279 x 19
##     year month   day dep_time sched_dep_time dep_delay arr_time
##    <int> <int> <int>    <int>          <int>     <dbl>    <int>
##  1  2013     1     1      542            540         2      923
##  2  2013     1     1      544            545        -1     1004
##  3  2013     1     1      557            600        -3      838
##  4  2013     1     1      558            600        -2      849
##  5  2013     1     1      558            600        -2      853
##  6  2013     1     1      558            600        -2      924
##  7  2013     1     1      559            559         0      702
##  8  2013     1     1      606            610        -4      837
##  9  2013     1     1      611            600        11      945
## 10  2013     1     1      613            610         3      925
## # ... with 111,269 more rows, and 12 more variables: sched_arr_time <int>,
## #   arr_delay <dbl>, carrier <chr>, flight <int>, tailnum <chr>,
## #   origin <chr>, dest <chr>, air_time <dbl>, distance <dbl>, hour <dbl>,
## #   minute <dbl>, time_hour <dttm>
## 
## [[3]]
## # A tibble: 104,662 x 19
##     year month   day dep_time sched_dep_time dep_delay arr_time
##    <int> <int> <int>    <int>          <int>     <dbl>    <int>
##  1  2013     1     1      533            529         4      850
##  2  2013     1     1      554            600        -6      812
##  3  2013     1     1      557            600        -3      709
##  4  2013     1     1      558            600        -2      753
##  5  2013     1     1      559            600        -1      941
##  6  2013     1     1      600            600         0      851
##  7  2013     1     1      600            600         0      837
##  8  2013     1     1      602            610        -8      812
##  9  2013     1     1      602            605        -3      821
## 10  2013     1     1      623            610        13      920
## # ... with 104,652 more rows, and 12 more variables: sched_arr_time <int>,
## #   arr_delay <dbl>, carrier <chr>, flight <int>, tailnum <chr>,
## #   origin <chr>, dest <chr>, air_time <dbl>, distance <dbl>, hour <dbl>,
## #   minute <dbl>, time_hour <dttm>

We can verify how the grouping was carried out by using group_keys:

group_keys(split_test)

## # A tibble: 3 x 1
##   origin
##   <chr> 
## 1 EWR   
## 2 JFK   
## 3 LGA

Now, let’s get rid of that, and try group_map, again, by origin. We may as well use my highly pointless function. If all goes well, we will return a tibble of 15 rows

rm(split_test)
test5 <- flights %>% 
  group_by(origin) %>% 
 group_map(~ header(.x))

The dplyr help tells us that ‘.x’ refers to the subset of rows for each group that is passed in. That returned 5 rows per origin, as expected.Let’s try again and obtain some more rows per group:

test10 <- flights %>% 
  group_by(origin) %>% 
 group_map(~ header(.x,10L))

Now we have 30 observations in our new tibble, so this all looks good. Can we split and map together?

#test_group_and_map <- flights %>% 
 # group_split(origin) %>% 
 #group_map(~ first5(.x,20))

No we cannot. Group_map works on grouped tibbles, and by splitting, we’re passing it a list. Instant fail. RTFM folks.

At this point, I can see group_map being pretty useful, but I will need to do a bit more work on the use case for group_split - at present, I’m not seeing what it offers over and above group_by, for staying in a purely dplyr workflow. I understand if you want to throw some purrr into the mix, it becomes more useful. As always, feel free to chip in if you have any thoughts or answers..I will try to play around with this more as well.

To leave a comment for the author, please follow the link and comment on their blog: HighlandR.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)