Selection with a filter on row number and value

user_9609758 0
Viewed 59

I have the following simple data.table "test". I would like to select all rows of row 3 to 8 with X equal to "A": 

library(data.table)
set.seed(1)
test <- data.table(X=c(rep("A",5),rep("B",5)),Y=rnorm(10),Z=rnorm(10))

test[3:8 & X == "A"] # gives the not desired output:

1: A -0.6264538  1.5117812
2: A  0.1836433  0.3898432
3: A -0.8356286 -0.6212406
4: A  1.5952808 -2.2146999
5: A  0.3295078  1.1249309
Warning message:
  In 3:8 & X == "A" :
  longer object length is not a multiple of shorter object length

# desired outcome:

3: A -0.8356286 -0.62124058
4: A  1.5952808 -2.21469989
5: A  0.3295078  1.12493092

Between row 3:8, I would like to select just the ones with X == "A". How is this possible? Please note that using test[3:8][X == "A"] seems not as an option, because I want to do some calculations on these rows which are saved in the original datatable.

data.table subset r
3 Answers

Here 3:8 is definitely not of the same length as the second expression (X == "A") and more over, we are compare a logical index with a numeric index. Instead, convert the first expression to logical by using %in% on the sequence of rows, then two things happen - 1) lengths become same, 2) Same type

test[(seq_len(.N) %in% 3:8) & X == "A"]
#    X          Y          Z
#1: A -0.8356286 -0.6212406
#2: A  1.5952808 -2.2146999
#3: A  0.3295078  1.1249309
user_37322710 
library(data.table)
set.seed(1)
test <- data.table(X=c(rep("A",5),rep("B",5)),Y=rnorm(10),Z=rnorm(10))

test[test[, .I %in% 3:8 & X == "A"], Z := Z+3][]
#>     X          Y           Z
#>  1: A -0.6264538  1.51178117
#>  2: A  0.1836433  0.38984324
#>  3: A -0.8356286  2.37875942
#>  4: A  1.5952808  0.78530011
#>  5: A  0.3295078  4.12493092
#>  6: B -0.8204684 -0.04493361
#>  7: B  0.4874291 -0.01619026
#>  8: B  0.7383247  0.94383621
#>  9: B  0.5757814  0.82122120
#> 10: B -0.3053884  0.59390132

Created on 2019-06-21 by the reprex package (v0.3.0)

user_64614620

If you need to select rows from certain index(3:8) and then filter out a variable with certain value(here X == 'A'), you can try out with `dplyr package':

library(data.table)
library(dplyr)
set.seed(1)
test <- data.table(X=c(rep("A",5),rep("B",5)),Y=rnorm(10),Z=rnorm(10))

test %>% slice(3:8) %>% filter(X == 'A')
user_63340610
Related