如何迭代 Python 中的词汇表？

Question

Here's what I mean: Say I have a list. Let's call it messages.

messages = ['hey how are you', 'doing good what about you']

My end goal is to run this list against another list of vocabulary, and if each word is in the vocab list, put it in another. This vocabulary list looks like this:

vocab = ['hey', 'how', 'you']

(Notice 'are' is omitted)
The final list of my formatted data right now looks like this:

final_list = np.array([['', '', '', ''], ['', '', '', '']])

I want it to look something like this:

final_list = np.array([['hey', 'how', 'you', ''], ['you', '', '', '']])

I have an idea using a for loop and enumerate() , but it's not working too well. Help would be appreciated!

Answer 1

Go over the list of messages. For each message, split it into words, take at most N (N=4) words, and pad with empty strings, if needed.

N = 4
data = []
for m in messages:
    words = [x for x in m.split() if x in vocab]
    data.append(words[:N] + (N - len(words)) * [""])
final_list = np.array(data)

For better performance, convert vocab to a set before the loop:

vocab = set(vocab)

Result:

array([['hey', 'how', 'you', ''],
       ['you', '', '', '']], dtype='<U3')

Answer 2

Try with two for loops:

vocab = ['hey', 'how', 'you']
messages = ['hey how are you', 'doing good what about you']
m = []
s = []
for x in messages:
  for y in x.split():
    if y in vocab:
      s.append(y)
  m.append(s)
  s = []
    
print(m)

To get empty elements:

vocab = ['hey', 'how', 'you']
messages = ['hey how are you', 'doing good what about you']
m = []
s = []
for x in messages:
  for y in x.split():
    if y in vocab:
      s.append(y)
    else:
      s.append('')
  m.append(s)
  s = []
    
print(m)

Answer 3

List comprehension is an efficient way to do this. You can then convert the output into an array if need be.

li = ['hey how are you', 'doing good what about you']
vocab = ['hey', 'how', 'you']

final_list = [[el if el in el2 else '' for el in vocab] for el2 in li]

print(final_list)

Output:

[['hey', 'how', 'you'], ['', '', 'you']]