R: Matching Strings stored in an array with strings of another array

Question

I have different Strings (one String may contain ~1-4 words) stored in a Large Character Object (38506 elements in total) and a set of 10 texts stored in a chr-Object (about 100 words each) that might contain one of the Strings from the Large Character Object.

Now I would like to extract possible matches from the text set for every String.

I already tried the following, with "a4" being the set of texts (chr-object) and "t" being the Large Character Object:

i = 1
 while(i < 38506){
    matches <- str_extract(a4, t[i])
    i <- i +1 }

However, after the operation, the object "matches" contains only 10 "NA"-elements, although there are definitely a few matching strings in some of the texts.

|| EDIT2:

Here's an reproducible example of what I am trying to do with x representing the Large Character Object and z representing the set of texts

Please note that the while-loop currently does not produce the outcome as displayed below, the example illustrates what the result should look like.

The actual resulting object contains only 8 NA - elements, so there must be some error in the loop, the str_extract or the pmax-function:

> x
[1] "Hey-ho!"              "This is"              "Just some random"    
[4] "text"                 "I am trying to match" "please help"         
[7] "very nice"                   "Thanks"  



z
[1] "My name is Thomas. This is my first project"
[2] "R is a cool tool"  
[3] "Hello, Hi There and Hey-ho!"
[4] "Can you please help me clean this mess?"    
[5] "All the best!" 
[6] "Is there a way to get to London by train?"


i <- 1
while(i < length(x)){
extraction <- str_extract(z, x[i])
resulting <- pmax(resulting, extraction)

i <- i + 1
}



> resulting
[1] "This is" NA   "Hey-ho"   "please help" NA   NA 

If someone wants to try exactly what I am doing, I have uploaded my actual data into a dropbox folder: https://www.dropbox.com/sh/2y7ogjxk1glddh1/AADrDveQguzChaaXXIeLfmIfa?dl=0

I read the files into R like this:

a4 <- readLines(file.path(".","a4.txt"))

t <- readLines(file.path(".","LargeCharacterObject.txt"))

Due to some formating issues the following replacements should be conducted before trying to match the strings:

a4 <- gsub('Ãœ', 'Ü', a4)
a4 <- gsub('Ãœ', 'Ü', a4)
a4 <- gsub('Ä', 'Ä', a4)
a4 <- gsub('ß', 'ß', a4)
a4 <- gsub('ä', 'ä', a4)
a4 <- gsub('ü', 'ü', a4)
a4 <- gsub('ö', 'ö', a4)
a4 <- gsub('Ö', 'Ö', a4)

t <- gsub('Ãœ', 'Ü', t)
t <- gsub('Ä', 'Ä', t)
t <- gsub('ß', 'ß', t)
t <- gsub('ä', 'ä', t)
t <- gsub('ü', 'ü', t)
t <- gsub('ö', 'ö', t)
t <- gsub('Ö', 'Ö', t)
t <- gsub('\\', '', t)

EDIT2 END ||

Do I somehow need to wrap t[i] in a Regex-pattern? Is this even feasible? Or am I using the wrong type of objects / the wrong extraction method?

I am grateful for any hints or ideas.

Thanks

EDIT

I forgot to mention earlier that the elements of the array should stay in the same order and contain also the elements without matches, so the result should look something like:

[1] "NA" "NA" "a" "NA" "b" "NA"

I already tried this:

i = 1
while(i < 38506){
 matches <- (str_extract(a4, t[i])
 result <- pmax(matches, result)
 i <- i +1}

But somehow "result" also contains only 10 "NA" elements after execution.

Answer 1

Putting aside other possible changes to your code, it's not doing as you expect because you are over-writing matches each time rather than appending to it.

Thus, this will likely work for you.

i = 1
while(i < 38506){
     matches <- c(matches, str_extract(a4, t[i]))
     i <- i +1
}

To demonstrate with a reproducible example, here is an analogy to what you are currently doing.

matches <- character()
for(l in letters){
    matches <- l
}
matches
# [1] "z"

This what you should be doing in this analogous example.

matches <- character()
for(l in letters){
    print(l)
    matches <- c(matches, l)
}
matches
# [1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j" "k" "l" "m" "n" "o" "p" "q" "r" "s" "t" "u" "v"
# [23] "w" "x" "y" "z"