I have a data set that looks something like this
ID <- c("BOB", "BBP", "GCG", "GGO", "GOP", "ICO", "BGP")
df <- data.frame(
actor = sample(ID, size = 5, replace = TRUE),
target = sample(ID, size = 5, replace = TRUE),
power = runif(5)
)
I want to run a for loop that will go through the data frame and create a new data frame for every unique ID that pulls all the rows where that ID appears in either "actor" or "target". So essentially I have a data frame where every instance of BOB is subsetted.
I will have 60+ individuals, so making an individual data frame first and then filling it from a for loop would be a ton of time.
Is there a way I can make a for loop where a data frame is made for every unique 3 letter code that subsets all of its instances from a larger dataset?
One option is dplyr::group_split() but that that's experimental so we'll stay away.
Though others might have better answers, here's what I did:
library(tidyverse)
set.seed(1)
ID <- c("BOB", "BBP", "GCG", "GGO", "GOP", "ICO", "BGP")
df <- tibble(
actor = sample(ID, size = 5, replace = TRUE),
target = sample(ID, size = 5, replace = TRUE),
power = runif(5)
) |>
mutate(row_num = row_number(), .before = 1) # to ensure rows aren't duplicated later
df
#> # A tibble: 5 × 4
#> row_num actor target power
#> <int> <chr> <chr> <dbl>
#> 1 1 BOB GOP 0.206
#> 2 2 GGO BGP 0.177
#> 3 3 BGP GCG 0.687
#> 4 4 BOB ICO 0.384
#> 5 5 BBP BBP 0.770
combined <- bind_rows(
df |> mutate(ID = actor),
df |> mutate(ID = target)
) |>
distinct(ID, row_num, .keep_all = TRUE) |> # so each row is only present once per ID
nest(.by = ID)
combined
#> # A tibble: 7 × 2
#> ID data
#> <chr> <list>
#> 1 BOB <tibble [2 × 4]>
#> 2 GGO <tibble [1 × 4]>
#> 3 BGP <tibble [2 × 4]>
#> 4 BBP <tibble [1 × 4]>
#> 5 GOP <tibble [1 × 4]>
#> 6 GCG <tibble [1 × 4]>
#> 7 ICO <tibble [1 × 4]>
combined |>
filter(ID == "BBP") |>
pull(data) |>
pluck(1)
#> # A tibble: 1 × 4
#> row_num actor target power
#> <int> <chr> <chr> <dbl>
#> 1 5 BBP BBP 0.770
Created on 2026-11-02 with reprex v2.1.1
try base R split(), works perfectly.