How to select rows that have text searching in all columns of a data frame

Question

I want to select only the lines that have the text "Try", similar to grep in Linux. Follow the example:

my.data <- data.frame(
  A = c("prot trypsina catalic", "7", "123", NA, "1419", "ab", "ab", "ab"),
  B = c("1416", "7", "123trypsina", "1011", "1416", "ab", "TRYPSIN", "ab"),
  c = c("b", "a", "trypsina123", "trypsin", "no", "ab", "ab", "ab"),
  d = seq(1:8),
  e = rep("please", 8))

The desired result would be as follows:

my.data[c(1,3,4,7),]
                      A           B           c d      e
1 prot trypsina catalic        1416           b 1 please
3                   123 123trypsina trypsina123 3 please
4                  <NA>        1011     trypsin 4 please
7                    ab     TRYPSIN          ab 7 please

Answer 1

A solution, in a line:

my.data[apply(my.data, 1, function(x) any(grepl("tryp", x, ignore.case = TRUE))), ]
#>                       A           B           c d      e
#> 1 prot trypsina catalic        1416           b 1 please
#> 3                   123 123trypsina trypsina123 3 please
#> 4                  <NA>        1011     trypsin 4 please
#> 7                    ab     TRYPSIN          ab 7 please

Explanation:

The function grepl returns a logical vector with the presence of the pattern; the option ignore.case is for case-independent search. For example, for column A:

grepl("tryp", my.data$B, ignore.case = TRUE)
#> [1] FALSE FALSE  TRUE FALSE FALSE FALSE  TRUE FALSE

any is used to see if there is any true occurrence, regardless of position. For example:

any(c(FALSE, FALSE, TRUE))
#> [1] TRUE

apply applies a function to a dimension of an arrangement; in the case of a data frame., 1 indicates to apply to lines. The result together with grepl and any is a logic vector indicating the lines that have some occurrence of the pattern:

apply(my.data, 1, function(x) any(grepl("tryp", x, ignore.case = TRUE)))
[1]  TRUE FALSE  TRUE  TRUE FALSE FALSE  TRUE FALSE

which is used to index lines from data frame., in the form df[vetorlogico, ]

Answer 2

A solution using functions of tidyverse is the following:

library(tidyverse)

my.data %>%
    filter_all(any_vars(str_detect(., pattern = "(?i)tryp")))
#>                       A           B           c d      e
#> 1 prot trypsina catalic        1416           b 1 please
#> 2                   123 123trypsina trypsina123 3 please
#> 3                  <NA>        1011     trypsin 4 please
#> 4                    ab     TRYPSIN          ab 7 please

^{Created on 2021-06-12 by the reprex package (v2.0.0)}

What the above command does is this:

• filter_all will filter all lines
• any_vars choose any column that satisfies a logical condition
• the logical condition is determined by str_detect, that will select any cell that contains the string tryp, whereas (?i) indicates that no differentiation is made between upper and lower case letters tryp