在列表中查找重复的子字符串

Question

我有一长串子字符串（接近 16000 个），我想找到重复循环开始/停止的位置。我想出了这段代码作为起点：

strings= ['1100100100000010',
        '1001001000000110',
        '0010010000001100',
        '0100100000011011',
        '1001000000110110',
        '0010000001101101',
        '1100100100000010',
        '1001001000000110',
        '0010010000001100',
        '0100100000011011',]

pat = [ '1100100100000010',
        '1001001000000110',
        '0010010000001100',]

for i in range(0,len(strings)-1):
    for j in range(0,len(pat)):
        if strings[i] == pat[j]:
            continue
        if strings[i+1] == pat[j]:
            print 'match', strings[i]
            break
        break

这种方法的问题是你必须知道 pat 是什么才能搜索它。我希望能够从第一个 n 子列表（在本例中为 3）开始并搜索它们，如果不匹配，则将一个子字符串向下移动到下一个 3，直到它遍历整个列表或找到重复。我相信如果长度足够长（可能是 10 个），它会找到重复而不需要太多时间。

Answer 1

strings= ['1100100100000010',
        '1001001000000110',
        '0010010000001100',
        '0100100000011011',
        '1001000000110110',
        '0010000001101101',
        '1100100100000010',
        '1001001000000110',
        '0010010000001100',
        '0100100000011011',]

n = 3

patt_dict = {}

for i in range(0, len(strings) - n, 1):
    patt = (' '.join(strings[i:i + n]))
    if patt not in patt_dict.keys(): patt_dict[patt] = 1
    else: patt_dict[patt] += 1


for key in patt_dict.keys():
    if patt_dict[key] > 1: 
        print 'Found ' + str(patt_dict[key]) + ' repeating instances of ' + str(key) + '.'

试一试。以线性时间运行。基本上使用字典来计算子集中出现 n 大小模式的次数。如果它超过 1，那么我们就有一个重复的模式 :)

Answer 2

这是一种相当简单的方法，可以找到所有长度 >= 1 的所有匹配项：

def findall(xs):
    from itertools import combinations
    # x2i maps each member of xs to a list of all the
    # indices at which that member appears.
    x2i = {}
    for i, x in enumerate(xs):
        x2i.setdefault(x, []).append(i)
    n = len(xs)
    for ixs in x2i.values():
        if len(ixs) > 1:
            for i, j in combinations(ixs, 2):
                length = 1 # xs[i] == xs[j]
                while (i + length < n and
                       j + length < n and
                       xs[i + length] == xs[j + length]):
                    length += 1
                yield i, j, length

然后：

for i, j, n in findall(strings):
    print("match of length", n, "at indices", i, "and", j)

显示：

match of length 4 at indices 0 and 6
match of length 1 at indices 3 and 9
match of length 3 at indices 1 and 7
match of length 2 at indices 2 and 8

您想要做什么和不想要什么并没有被精确指定，所以这里列出了 所有 匹配项。你可能真的不想要其中的一些。例如，索引 1 和 7 处长度为 3 的匹配只是索引 0 和 6 处长度为 4 的匹配的尾端。

因此，您需要更改代码来计算您真正想要的。也许您只想要一个最大的匹配？所有最大匹配？只有特定长度的匹配？等等。

Answer 3

这里可以找到字符串数组中所有匹配的子数组。

strings = ['A', 'B', 'C', 'D', 'Z', 'B', 'B', 'C', 'A', 'B', 'C']

pat = ['A', 'B', 'C', 'D']

i = 0
while i < len(strings):
    if strings[i] not in pat:
        i += 1
        continue
    matches = 0
    for j in xrange(pat.index(strings[i]), len(pat)):
        if i + j - pat.index(strings[i]) >= len(strings):
            break
        if strings[i + j - pat.index(strings[i])] == pat[j]:
            matches += 1
        else:
            break
    if matches:
        print 'matched at index %d subsequence length: %d value %s' % (i, matches, strings[i])
        i += matches
    else:
        i += 1

输出：

matched at index 0 subsequence length: 4 value A
matched at index 5 subsequence length: 1 value B
matched at index 6 subsequence length: 2 value B
matched at index 8 subsequence length: 3 value A