Also of note is the hashslice with the null assign; this is actually a bit of an obfuscated way to remove all duplicated non-stopwords in the data. I wouldn't recommend using this method; its probably not very efficient, and its definately not very obvious.