Why do I get an IndexError?

Question

I am trying to get trigrams out of a sentence and save them in a dictionary, with their frequenz as value. I wrote this:

trigrams = {}
sentence = ["What", "is", "happening", "right", "now"]

for word in sentence:
      if word != sentence[-1] or sentence[-2] and tuple((word, sentence[sentence.index(word) +1], sentence[sentence.index(word) +2])) not in trigrams:
             trigrams.update({tuple((word, sentence[sentence.index(word) +1], sentence[sentence.index(word) +2])):1})

Should look like this: ("what","is","happening"):1 ("is","happening","right"):1 etc

But now I am keep getting an IndexError in the update-line.

`word != sentence[-1] or sentence[-2]`: that is not what you wanted to do. — Jean-François Fabre, Dec 18 '16 at 17:29
I can't build trigrams with the last two words as the fist word (right, now, ???) so i wont do anything with them. Therefore the test if the current word is one of the last two words. — spiderkitty, Dec 18 '16 at 17:50

score 0 · Answer 1 · answered Dec 18 '16 at 17:31

0

You can use lists as your tuples' contents are all of the same datatype (string)

It's probably easier to do:

trigrams = []
sentence = ["What", "is", "happening", "right", "now"]

for i in range(2,len(sentence)):
    trigrams.append([sentence[i-2],sentence[i-1],sentence[i]])

answered Dec 18 '16 at 17:31

Ben Stobbs

422
6
14

Yes, that actually looks a lot easier but I need to test, if they are alredy in the dict. However, I found my mistake. Thank you for helping me! – spiderkitty Dec 18 '16 at 18:02

score 0 · Accepted Answer · answered Dec 18 '16 at 17:32

0

I guess if word != sentence[-1] or sentence[-2] is not what you want. Do you mean if word != sentence[-1] and word != sentence[-2], meaning word does not equal either sentence[-1] nor sentence[-2]?

answered Dec 18 '16 at 17:32

jmd_dk

12,125
9
63
94

Oh yeah, that was actually causing the problem :D Thank you so much, it is working now! – spiderkitty Dec 18 '16 at 17:58

fedepad · Answer 3 · 2016-12-18T17:58:17.960

Given you would like to keep your code structure with the tuple and change minimally your code, you can do this (not saying this might be a good approach for your problem, etc.):

trigrams = {}
sentence = ["What", "is", "happening", "right", "now"]

for index, word in enumerate(sentence):
    print index, word  # to understand how the iteration goes on
    if index < len(sentence)-2:
        if tuple((word, sentence[index+1], sentence[index+2])) not in trigrams:
            trigrams.update({tuple((word, sentence[index+1], sentence[index+2])):1})

You were getting an index error because you were accessing an element that didn't exist in tuple()... because the way you were doing the checking to see if were near the end of the list (the last two elements) wasn't done right.

The code you were using:

if word != sentence[-1] or sentence[-2]

is not right and you were comparing strings eventually and not the indexes, which is what is important here! Compare the indexes, not the values at those positions.

Yes, it worked as I replaced the "or" with an "and". Thank you for helping me! — spiderkitty, Dec 18 '16 at 18:00

Why do I get an IndexError?

3 Answers3