从列表中删除元组会删除一些但不是全部

Question

I must be missing something really obvious.

I have a list of tuples that are (phrase, number) pairs. I want to remove entire tuples that have phrases containing stopwords from my stopwords list.

stopwords = ['for', 'with', 'and', 'in', 'on', 'down']
tup_list = [('faucet', 5185), ('kitchen', 2719), ('faucets', 2628),
            ('kitchen faucet', 1511), ('shower', 1471), ('bathroom', 1131),
            ('handle', 1048), ('for', 1035), ('cheap', 960), ('bronze', 807),
            ('tub', 797), ('sale', 771), ('sink', 762), ('with', 696),
            ('single', 620), ('kitchen faucets', 615), ('stainless faucet', 613),
            ('pull', 603), ('and', 477), ('in', 447), ('single handle', 430),
            ('for sale', 406), ('bathroom faucet', 392), ('on', 369),
            ('down', 363), ('head', 359), ('pull down', 357), ('wall', 351),
            ('faucet with', 350)]

for p,n in tup_list:
    print('p', p, p.split(), any(phrase in stopwords for phrase in p.split()))

print(len(tup_list))
for p,n in tup_list:
    if any(phrase in stopwords for phrase in p.split()):
        tup_list.remove((p,n))
        print('Removing', p)
print(len(tup_list))

print([item for item in tup_list if item[0] == 'in'])

When I run the above, I get the following print-out:

p faucet ['faucet'] False
p kitchen ['kitchen'] False
p faucets ['faucets'] False
p kitchen faucet ['kitchen', 'faucet'] False
p shower ['shower'] False
p bathroom ['bathroom'] False
p handle ['handle'] False
p for ['for'] True
p cheap ['cheap'] False
p bronze ['bronze'] False
p tub ['tub'] False
p sale ['sale'] False
p sink ['sink'] False
p with ['with'] True
p single ['single'] False
p kitchen faucets ['kitchen', 'faucets'] False
p stainless faucet ['stainless', 'faucet'] False
p pull ['pull'] False
p and ['and'] True
p in ['in'] True
p single handle ['single', 'handle'] False
p for sale ['for', 'sale'] True
p bathroom faucet ['bathroom', 'faucet'] False
p on ['on'] True
p down ['down'] True
p head ['head'] False
p pull down ['pull', 'down'] True
p wall ['wall'] False
p faucet with ['faucet', 'with'] True
29
Removing for
Removing with
Removing and
Removing for sale
Removing on
Removing pull down
Removing faucet with
22
[('in', 447)]

My Question : why doesn't the tuple containing ('in', 447) get removed? The printout shows p in ['in'] True meaning 'in' is in the stopwords list, so why does tup_list.remove((p,n)) not remove it?

Answer 1

When you remove an item from the list in place, the indices change. As you iterate over a list that changes, you will see unexpected results.

Here is one solution. It's not the most efficient one, but may suit your needs.

remove_indices = []

for i, (p, n) in enumerate(tup_list):
    if any(phrase in stopwords for phrase in p.split()):
        remove_indices.append(i)
        print('Removing', p)

tup_list = [i for j, i in enumerate(tup_list) if j not in remove_indices]