Parse a character string and later re-assemble it - r

Question

I am trying to parse a character string into its parts, check if each of the parts exist in a separate vocabulary, and later re-assemble only those strings whose parts are in the vocabulary. The vocabulary is a vector of words, and is created separately from the the strings I want to compare. The final goal is to create a data frame with only those strings whose word parts are in the vocabulary.
I have written a piece of code to parse out the data into strings, but cannot figure out how to make the comparison. If you believe that parsing out the data is not the optimal solution, please let me know.
Here is an example:
Assume that I have three character strings:
"The elephant in the room is blue",
"The dog cannot swim",
"The cat is blue"
and my vocabulary consists of the words:
cat, **the**, **elephant**, hippo,
**in**, run, **is**, bike,
walk, **room, is, blue, cannot**
In this case I will pick only the first and third strings, because each of their word parts are matched to a word in my vocabulary. I will not select the second string, because the words "dog" and "swim" are not in the vocabulary.
Thank you!
Per request, attached is the code I have written so far to clean the strings, and parse them into unique words:
animals <- c("The elephant in the room is blue", "The dog cannot swim", "The cat is blue")
animals2 <- toupper(animals)
animals2 <- gsub("[[:punct:]]", " ", animals2)
animals2 <- gsub("(^ +)|( +$)|( +)", " ", animals2)
## Parse the characters and select unique words only
animals2 <- unlist(strsplit(animals2," "))
animals2 <- unique(animals2)

Answer 1

Here how I would do :
Read the data
clean vocab to remove extra spaces and *
Loop over strings , using setdiff
My code is :
## read your data
tt <- c("The elephant in the room is blue",
"The dog cannot swim",
"The cat is blue")
vocab <- scan(textConnection('cat, **the**, **elephant**, hippo,
**in**, run, **is**, bike,
walk, **room, is, blue, cannot**'),sep=',',what='char')
## polish vocab
vocab <- gsub('\\s+|[*]+','',vocab)
vocab <- vocab[nchar(vocab) >0]
##
sapply(tt,function(x){
+ x.words <- tolower(unlist(strsplit(x,' '))) ## take lower (the==The)
+ length(setdiff(x.words ,vocab)) ==0
+ })
The elephant in the room is blue The dog cannot swim The cat is blue
TRUE FALSE TRUE