Count number of multiple pattern matches in a string

Question

I would like to count multiple pattern matches in a dataframe column containing long strings.

pattern<-c("AAA", "BBB", "CCC")

df$AAA <- str_count(df$string_1, "AAA+")
df$BBB <- str_count(df$string_1, "BBB+")
df$CCC <- str_count(df$string_1, "CCC+")
df$AAA <- str_count(df$string_2, "AAA+")
df$BBB <- str_count(df$string_2, "BBB+")
df$CCC <- str_count(df$string_2, "CCC+")
...

In reality the list "pattern" is much longer, so need to use loop over the column with strings.

score 3 · Accepted Answer · answered Aug 15 '17 at 20:03

3

You could loop with lapply or sapply

#DATA
pattern<-c("AAA", "BBB", "CCC")
set.seed(42)
df = data.frame(a = replicate(5, paste(sample(c("A", "B", "C"), 50, TRUE), collapse = "")),
                b = replicate(5, paste(sample(c("A", "B", "C"), 50, TRUE), collapse = "")))

library(stringr)
setNames(lapply(pattern, function(x) sapply(df, function(y)
                              str_count(string = y, pattern = x))), pattern)
#$AAA
#     a b
#[1,] 0 0
#[2,] 2 1
#[3,] 0 2
#[4,] 4 1
#[5,] 2 2

#$BBB
#     a b
#[1,] 1 2
#[2,] 1 0
#[3,] 2 3
#[4,] 1 2
#[5,] 2 1

#$CCC
#     a b
#[1,] 1 0
#[2,] 2 1
#[3,] 2 0
#[4,] 2 0
#[5,] 0 1

answered Aug 15 '17 at 20:03

d.b

32,245
6
36
77

getting this error: ' Error in UseMethod("type") : no applicable method for 'type' applied to an object of class "factor" ' – user2904120 Aug 15 '17 at 20:37
1

yes, it helped, both pattern list and table column, as.character – user2904120 Aug 15 '17 at 20:47
but the solution results in multiple matrices; also if the table contains more columns, each matrix contains all of them – user2904120 Aug 15 '17 at 20:48
i would like it to check only against the 'string' column, where strings are – user2904120 Aug 15 '17 at 20:57
yes, sth like this: setNames(lapply(pattern, function(x) sapply(df[,c("a")], function(y) str_count(string = y, pattern = x))), pattern) – user2904120 Aug 15 '17 at 21:17
and how to place all the data in columns of a new data frame, I mean AAA$a, BBB$a, CCC$a – user2904120 Aug 15 '17 at 21:36
great solution, it all worked, i'm just not able to fully understand it, would be great with some classical for loops – user2904120 Aug 16 '17 at 08:26

Count number of multiple pattern matches in a string

1 Answers1