how to remove single character(letter) from corpus using python

Question

I would like to remove any single characters from each document in corpus. For example, let's say there are some typos or non-English letters.

corpus=['I like this d hotel room because it was clean.','This hotel is very y close to downtown area.']

What I've tried was

corpus=' '.join( [w for w in corpus.split() if len(w)>1] )

but didn't work. Could anyone help me out?

Answer 1

Try the below

corpus = ['I like this d hotel room because it was clean.', 'This hotel is very y close to downtown area.']
corpus1 = []
for entry in corpus:
    corpus1.append(' '.join(x for x in entry.split() if len(x) > 1))
print(corpus1)

output

['like this hotel room because it was clean.', 'This hotel is very close to downtown area.']

Answer 2

This should work for you:

corpus=['I like this d hotel room because it was clean.','This hotel is very y close to downtown area.']
clean_corpus=[]
for sentence in corpus:
    clean_sentence=[]
    parts=sentence.split(" ")
    for part in parts:
        invalid=False
        if (len(part)==1) and (part.lower()!="a") and (part.lower()!="i") and (not part.isdigit()):
            invalid=True
        if not invalid:
            clean_sentence.append(part)
    clean_corpus.append(" ".join(clean_sentence))
print(clean_corpus)

This cleans out all single letter words that are not "a", "A", "i", "I", or a digit (1, 2, 3, ...).

Try it yourself and tell me in the comments if it worked or what could be improved!