I want to remove stopwords from text but fail to use regex and variables properly. For example I remove the stopword "he" but this also affects the word "when". I tried to use word boundaries like this:
new RegExp('\b'+stopwords[i]+'\b' , 'g') but doesn't work...
See a small example here: jsFiddle
var stopwords = ['as', 'at', 'he', 'the', 'was'];
for (i = 0; i < stopwords.length; i++) {
str = str.replace(new RegExp(stopwords[i], 'g'), '');
}
Something like this maybe
str = str.replace(new RegExp('\\b('+stopwords.join('|')+')\\b', 'g'), '');
You have to double escape in RegExp, and you could just join everything creating
/\b(as|at|he|the|was)\b/g
for loop? Consider that I have a list of 300 stopwords.
.join('|') Sorry but I noticed this difference now :-)
\b anchors to work properly, i.e. the pattern should be: /\b(?:word1|word2)\b/ instead of: /\bword1|word2\b/.join('|') joins an array with the pipe as "glue", and as noted by ridgerunner there's parenthesis added to create the regex I posted in the answer.Use \\b to make a single \b.
new RegExp('\\b'+stopwords[i]+'\\b' , 'g')
You need to escape backslash becasue it's inside string literal, not in the regular expression:
new RegExp('\\b' + stopwords[i] + '\\b' , 'g')
Otherwise, '\b' is BACKSPACE character ('\x08').
