在 R 中，如何只删除满足特定条件的特定百分比的行？

Question

I have a dataset that has 40% females and 60% males. I also have a column for job role, where 85% are management. I want to drop rows randomy until I reach a maximum of 50% males and 50% with the job role management.

I can find several solutions for how to drop all rows that meet those conditions, but nothing that lets me specifiy only dropping a certain number or percentage of rows.

Can anyone suggest code that would achieve this?

Answer 1

Starting with some fake data:

set.seed(42)
df1 <- data.frame(gender = sample(c("M", "F"), 1000, replace = TRUE, c(0.4, 0.6)),
                  role = sample(c("mgmt", "other"), 1000, replace = TRUE, c(0.85, 0.15)))

prop.table(table(df1))

#      role
#gender  mgmt other
#     F 0.529 0.094
#     M 0.324 0.053

We could look at the existing proportions and then sample using the ratio of what we want to what we have:

library(dplyr)
props <- df1 %>%
  count(gender, role) %>%
  mutate(share = n / sum(n),
         desired = 0.25,
         weighting = desired/share)

df2 <- df1 %>%
  left_join(props) %>%
  slice_sample(n = 100, weight_by = weighting) %>%
  select(gender, role)

prop.table(table(df2))


      role
gender mgmt other
     F 0.23  0.22
     M 0.31  0.24