phplyr

Data & Workflow

Statistics, Modeling, & Visualization

Pivot between wide and long

Count the penguins per species and study year, with one column per year, then stack the year columns into one row per species and year. penguins has size measurements for 344 penguins near Palmer Station, Antarctica, from 2007 to 2009 (Gorman et al., 2014).

use Phibble\DataFrame;use function Phplyr\{count, pivot_longer, pivot_wider};use function Phplyr\starts_with;$data = 'https://raw.githubusercontent.com/phpstats'    . '/datasets/main/csv';$penguins = DataFrame::fromCsv("$data/penguins.csv");$wide = $penguins |>    count('species', 'year') |>    pivot_wider(        namesFrom: 'year',        valuesFrom: 'n',        namesPrefix: 'n_',    );echo $wide, "\n\n";echo $wide |> pivot_longer(    starts_with('n_'),    namesTo: 'year',    namesPrefix: 'n_',    valuesTo: 'penguins',);
use Phibble\DataFrame;use function Phplyr\{phplyr, starts_with};$data = 'https://raw.githubusercontent.com/phpstats'    . '/datasets/main/csv';$penguins = DataFrame::fromCsv("$data/penguins.csv");$wide = phplyr($penguins)    ->count('species', 'year')    ->pivotWider(        namesFrom: 'year',        valuesFrom: 'n',        namesPrefix: 'n_',    )    ->get();echo $wide, "\n\n";echo phplyr($wide)->pivotLonger(    starts_with('n_'),    namesTo: 'year',    namesPrefix: 'n_',    valuesTo: 'penguins',);
# A tibble: 3 × 4
  species   n_2007 n_2008 n_2009
  <chr>      <int>  <int>  <int>
1 Adelie        50     50     52
2 Chinstrap     26     18     24
3 Gentoo        34     46     44

# A tibble: 9 × 3
  species   year  penguins
  <chr>     <chr>    <int>
1 Adelie    2007        50
2 Adelie    2008        50
3 Adelie    2009        52
4 Chinstrap 2007        26
5 Chinstrap 2008        18
6 Chinstrap 2009        24
7 Gentoo    2007        34
8 Gentoo    2008        46
9 Gentoo    2009        44
  • count() gives one row per species and year, with the count in n.
  • pivot_wider() makes one column for each value of year, adds the n_ prefix, and fills the columns from n. The columns it doesn’t use, here species, identify the output rows.
  • pivot_longer() stacks the columns that starts_with('n_') selects. Each column name, without the n_ prefix, goes into year, and each value goes into penguins.
  • year holds strings, because the values come from column names.

Reference: tidyr · Selecting columns