Filter for two conditions in R and combine in one data table-CodePudding

I have the following table:

totalAdjustments	accuracy
1	1
2	2
4	5
1	3

And would like to group it by totalAdjustments into two groups:

group1: totalAdjustments == 1 (named: oneAdjustment)

group2: totalAdjustments >= 2 (named: twoOrMoreAdjustments)

To get the following table:

numberOfAdjustments	accuracy
oneAdjustment	1
twoOrMoreAdjustments	2
twoOrMoreAdjustments	5
oneAdjustment	3

I currently import my csv with fread

result <- fread("data.csv")

CodePudding user response：

base R

You can use ifelse for that:

ifelse(dat$totalAdjustments > 1, "twoOrMore", "one")
# [1] "one"       "twoOrMore" "twoOrMore" "one"      
dat$totalAdjustments <- ifelse(dat$totalAdjustments > 1, "twoOrMore", "one")
dat
#   totalAdjustments accuracy
# 1              one        1
# 2        twoOrMore        2
# 3        twoOrMore        5
# 4              one        3

dplyr

library(dplyr)
dat %>%
  mutate(totalAdjustments = if_else(totalAdjustments > 1, "twoOrMore", "one"))
#   totalAdjustments accuracy
# 1              one        1
# 2        twoOrMore        2
# 3        twoOrMore        5
# 4              one        3

If this is expanded to include another number, perhaps

greater than 3 --> "tooMany"

then I would shift from a simple ifelse flow to cut:

dat %>%
  mutate(totalAdjustments = cut(totalAdjustments, c(0, 1, 3, Inf), c("one", "twoOrMore", "tooMany")))
#   totalAdjustments accuracy
# 1              one        1
# 2        twoOrMore        2
# 3          tooMany        5
# 4              one        3

Note that totalAdjustments is now of class factor instead of character; the difference may be nothing, but often it can lead to unexpected results if you do not intend the class; in that case, wrap it with as.character, as in = as.character(cut(...)).