I have a data frame containing following columns:-
sample.data
a_b_c d_b_e r_f_g c_b_a
1 1 1 1 1
2 2 2 2 2
3 3 3 3 3
4 4 4 4 4
How do I select only columns that contain both let's say "a" and "c" in the column name?
To select variables that contain a and c we could do:
library(dplyr)
df %>%
select(matches("(a.*c)|(c.*a)"))
a_b_c c_b_a 1 1 1 2 2 2 3 3 3 4 4 4
Note that var a_a_e is not selected because it doesn't contain c and var c_f_g is not selected because it doesn't contain a. Column names with two a's and two c's will not be selected either as seen with var a_a_e.
We could also use str_subset:
library(dplyr)
library(stringr)
df %>%
select(str_subset(names(df), "(a.*c)|(c.*a)"))
Data:
df <- data.frame(
a_b_c = 1:4,
a_a_e = 1:4,
c_f_g = 1:4,
c_b_a = 1:4
)
Try df %>% dplyr::select(matches("(a|c)"))
library(dplyr)
df <- data.frame(
a_b_c=1:4,
d_b_e=1:4,
r_f_g=1:4,
c_b_a=1:4
)
Results
> df %>% dplyr::select(matches("(a|c)"))
a_b_c c_b_a
1 1 1
2 2 2
3 3 3
4 4 4
df %>% select(intersect(contains("a"), contains("c"))) in that case.If you want to see how it works under the hood, use the following function:
contain_both <- function(data_frame, letter_a, letter_b) {
j <- 0
keep_columns <- NULL
for(i in 1:ncol(data_frame)) {
has_letters <- unlist(strsplit(names(data_frame)[i], '_'))
if(is.element(letter_a, has_letters) && is.element(letter_b, has_letters)) {
j <- j + 1
keep_columns[j] <- i
}
}
return(data_frame[, keep_columns])
}
Data:
df <- data.frame(seq(1:4), seq(1:4), seq(1:4), seq(1:4))
names(df) <- c('a_b_c', 'd_b_e', 'r_f_g', 'c_b_a')
Just pass in your data frame, along with your 2 letter choices:
Usage:
contain_both(df, 'b', 'c')
Hope this is what you are looking for:
a_b_c <- c(1,2,3,4)
d_b_e <- c(1,2,3,4)
yy <- cbind(a_b_c, d_b_e)
> yy
a_b_c d_b_e
[1,] 1 1
[2,] 2 2
[3,] 3 3
[4,] 4 4
yy <- as.data.frame(yy)
yy
a_b_c d_b_e
1 1 1
2 2 2
3 3 3
4 4 4
y <- yy[which(names(yy) %in% "a_b_c")]
> y
a_b_c
1 1
2 2
3 3
4 4
In your example, you can use this:
y <- sample.data[which(names(sample.data) %in% c("a_b_c","c_b_a" )]