"

In today’s #rstats adventure I learnt about tidyr::complete()

r-how-to
Running summarise() or count() and want to keep combinations where there are 0 entries? complete() is your friend!
Author
Affiliation

Cara Thompson

Building Stories with Data

Published

May 26, 2022

Quick demo 👇

First, the default behaviour.

R script output showing a tibble with 5 rows and 3 columns. Groups penguin species by island. Adelie penguins on Biscoe, Dream, and Torgersen islands number 44, 56, and 52, respectively. Chinstrap on Dream has 68, Gentoo on Biscoe has 124.

palmerpenguins::penguins %>%
   group_by(species, island) %>%
   count()

# A tibble: 5 x 3
# Groups:   species, island [5]
  species   island        n
  <fct>     <fct>     <int>
1 Adelie    Biscoe       44
2 Adelie    Dream        56
3 Adelie    Torgersen    52
4 Chinstrap Dream        68
5 Gentoo    Biscoe      124

Next, a surprise.

If we just add complete() into the sequence, we end up with 45 rows! There are only 3 penguin species and 3 islands, so we’d only expect 9.

See the note above the table? We have 5 groups in the existing data. 45 = 9 * 5. Aha!

R script: palmerpenguins::penguins %>% group_by(species, island) %>% count() %>% complete(species, island)

We need to ungroup() before running complete()

Otherwise we’re completing within each current grouping, which isn’t what we want to do.

Tada!

R script: palmerpenguins::penguins %>% group_by(species, island) %>% count() %>% ungroup() %>% complete(species, island); then a tibble displaying penguin species distribution across islands with missing data. Columns: species, island, count 'n'. Adélie has counts for all islands, Chinstrap lacks data for two, and Gentoo is only recorded on Biscoe.

But let’s be more specific

We started this exercise because we wanted to be explicit about there being no penguins of some species on some of the islands so we want to replace the NAs with 0s.

To do that, we need to provide a list of values to swap in for NAs in each column we need to fill. Here that’s n. 

R script same as above, but in which we've replaced the final line with: complete(species, island, fill = list(n = 0)); A tibble of penguin species counts across three islands: Biscoe, Dream, and Torgersen.

A word of warning

If zeros aren’t what you’re after, choose something else. Just beware of that that will do that the class of your column!

R script same as above, but in which we've replaced the final line with: complete(species, island, fill = list(n = 'No such penguins here')); in the tibble, the column counting penguins has now been turned into character.

Reuse

Citation

For attribution, please cite this work as:
Thompson, Cara. 2022. “In Today’s #Rstats Adventure I Learnt about Tidyr::complete().” May 26, 2022. https://www.cararthompson.com/posts/2022-05-26-in-todays-rstats-adventure-i/.