CODEWITHSUNDEEP
rdata.tablesubset
Arpit
Arpit

Reputation: 53

Subset data.table by one variable based on conditional values in other variables

I have a data.table in R that I need to subset based on certain conditions

dt <- data.table(X_ID=c(1,1,1,1,1,1,2,2,2,2,2,2,3,3,3,3,3,3,4,4,4,4,4,4), Y_ID=c(1,2,3,4,5,6, 1,2,3,4,5,6, 1,2,3,4,5,6, 1,2,3,4,5,6), Value1=c(6,4,8,9,10,12,5,7,8,3,4,8,7,3,9,7,4,12,45,3,2,1,2,5),Value2=c(6,6,8,9,10,12,5,4,8,3,4,8,7,5,9,7,4,12,45,1,2,1,2,5))

I want to retain all X_ID's in which value 1 is greater than value 2 when Y_ID is equal to 2. E.g., for X_ID=1, value 1 is greater than value 2 when Y_ID=2. In this case, I want to retain all rows that have X_ID=1.

I am currently using the filter function (shown below) to get at the solution. 

dt2 <- filter(dt,is.na(match(dt$X_ID,filter(dt, Y_ID==2&Value1<Value2)$X_ID)))

But is there a way to make this code more readable? Or use something native to data.table to subset the rows I require?

Upvotes: 3

Views: 939

Answers (2)

LocoGris
LocoGris

Reputation: 4480

Using data.table:

    dt<-dt[X_ID %in% dt[(Y_ID == 2 & (Value1 > Value2))]$X_ID]

#    X_ID Y_ID Value1 Value2
#1:    2    1      5      5
#2:    2    2      7      4
#3:    2    3      8      8
#4:    2    4      3      3
#5:    2    5      4      4
#6:    2    6      8      8
#7:    4    1     45     45
#8:    4    2      3      1
#9:    4    3      2      2
#10:    4    4      1      1
#11:    4    5      2      2
#12:    4    6      5      5

Upvotes: 3

Ronak Shah
Ronak Shah

Reputation: 389047

Using dplyr we can group_by X_ID and filter for groups where there is at least one value for Y_ID == 2 and Value1 > Value2. 

library(dplyr)

dt %>%
  group_by(X_ID) %>%
  filter(any(Y_ID == 2 & Value1 > Value2))

#    X_ID  Y_ID Value1 Value2
#   <dbl> <dbl>  <dbl>  <dbl>
# 1     2     1      5      5
# 2     2     2      7      4
# 3     2     3      8      8
# 4     2     4      3      3
# 5     2     5      4      4
# 6     2     6      8      8
# 7     4     1     45     45
# 8     4     2      3      1
# 9     4     3      2      2
#10     4     4      1      1
#11     4     5      2      2
#12     4     6      5      5

Upvotes: 2

Related Questions