CODEWITHSUNDEEP
rdplyr
ana_gg
ana_gg

Reputation: 370

Change column based on other two columns

I have a dataframe as follows:

df <- data.frame(matrix(NA, nrow = 5, ncol = 1))
colnames(df) <- "father"
df$father <- c("A", "B", "B","C","D")
df$id <- c("C", "E","F","G","H")
df$pop <- c("ref","ref","ref","val","val")

Which gives:

  father id pop
1      A  C ref
2      B  E ref
3      B  F ref
4      C  G val
5      D  H val

Then if "id" appears in "father", the value of the pop should be the one that appears in the row of the father. For example in this example "C" appears both in "father" and in "id", in father C the pop is val, then, I want the pop of C as id to be val as follows:

  father id pop
1      A  C val 
2      B  E ref
3      B  F ref
4      C  G val
5      D  H val

Any ideas?

Upvotes: 2

Views: 67

Answers (1)

runr
runr

Reputation: 1146

You can try the following

require(dplyr)
require(magrittr)

merge(x = df,
      y = df,
      by.x = 'id',
      by.y = 'father',
      all.x = T) %>% 
  mutate(., pop = ifelse(is.na(pop.y), pop.x, pop.y)) %>%
  subset(., select = c('father', 'id', 'pop'))

The idea is to perform a left join onto itself by merge (since all.x=T), and select the wanted value by mutate.

Upvotes: 1

Related Questions