如何在文本文件上执行二进制搜索以在python中搜索关键字?

时间:2011-03-07 08:56:22

标签: python search binary text-files

文本文件包含两列 - 索引号(5个空格)和字符(30个空格)。 它按字典顺序排列。我想执行二进制搜索来搜索关键字。

6 个答案:

答案 0 :(得分:9)

使用Python的内置bisect模块,这是一种有趣的方法。

import bisect
import os


class Query(object):

    def __init__(self, query, index=5):
        self.query = query
        self.index = index

    def __lt__(self, comparable):
        return self.query < comparable[self.index:]


class FileSearcher(object):

    def __init__(self, file_pointer, record_size=35):
        self.file_pointer = file_pointer
        self.file_pointer.seek(0, os.SEEK_END)
        self.record_size = record_size + len(os.linesep)
        self.num_bytes = self.file_pointer.tell()
        self.file_size = (self.num_bytes // self.record_size)

    def __len__(self):
        return self.file_size

    def __getitem__(self, item):
        self.file_pointer.seek(item * self.record_size)
        return self.file_pointer.read(self.record_size)


if __name__ == '__main__':
    with open('data.dat') as file_to_search:
        query = raw_input('Query: ')
        wrapped_query = Query(query)

        searchable_file = FileSearcher(file_to_search)
        print "Located @ line: ", bisect.bisect(searchable_file, wrapped_query)

答案 1 :(得分:3)

你需要做二分搜索吗?如果没有,请尝试将平面文件转换为cdb (constant database)。这将为您提供非常快速的哈希查找,以查找给定单词的索引:

import cdb

# convert the corpus file to a constant database one time
db = cdb.cdbmake('corpus.db', 'corpus.db_temp')
for line in open('largecorpus.txt', 'r'):
    index, word = line.split()
    db.add(word, index)
db.finish()

在单独的脚本中,针对它运行查询:

import cdb
db = cdb.init('corpus.db')
db.get('chaos')
12345

答案 2 :(得分:1)

如果您需要在文件中找到单个关键字:

line_with_keyword = next((line for line in open('file') if keyword in line),None)
if line_with_keyword is not None: 
   print line_with_keyword # found

要查找多个关键字,您可以将set()用作@kriegar suggested

def extract_keyword(line):
    return line[5:35] # assuming keyword starts on 6 position and has length 30

with open('file') as f:
    keywords = set(extract_keyword(line) for line in f) # O(n) creation
    if keyword in keywords: # O(1) search
       print(keyword)

您可以使用上面的dict()代替set()来保存index信息。

以下是如何对文本文件进行二进制搜索:

import bisect

lines = open('file').readlines() # O(n) list creation
keywords = map(extract_keyword, lines) 
i = bisect.bisect_left(keywords, keyword) # O(log(n)) search
if keyword == keywords[i]:
   print(lines[i]) # found

set()变体相比没有优势。

注意:除第一个之外的所有变体都将整个文件加载到内存中。 FileSearcher() suggested by @Mahmoud Abdelkader不需要将整个文件加载到内存中。

答案 3 :(得分:1)

我写了一个简单的Python 3.6+包就可以做到这一点。 (有关更多信息,请参见其github页!)

安装:pip install binary_file_search

示例文件:

1,one
2,two_a
2,two_b
3,three

用法:

from binary_file_search.BinaryFileSearch import BinaryFileSearch
with BinaryFileSearch('example.file', sep=',', string_mode=False) as bfs:
    # assert bfs.is_file_sorted()  # test if the file is sorted.
    print(bfs.search(2))

结果:[[2, 'two_a'], [2, 'two_b']]

答案 4 :(得分:0)

考虑使用集合代替二进制搜索来查找文件中的关键字。

集:

O(n)创建,O(1)查找,O(1)插入/删除

如果您的输入文件以空格分隔,则:

f = open('file')
keywords = set( (line.strip().split(" ")[1] for line in f.readlines()) )
f.close()    

my_word in keywords
<returns True or False>

字典:

f = open('file')
keywords = dict( [ (pair[1],pair[0]) for pair in  [line.strip().split(" ") for line in f.readlines()] ] ) 
f.close()

keywords[my_word]
<returns index of my_word>

二进制搜索是:

O(n log n)create,O(log n)lookup

编辑:对于5个字符和30个字符的情况,您可以使用字符串切片

f = open('file')
keywords = set( (line[5:-1] for line in f.readlines()) )
f.close()

myword_ in keywords

or 

f = open('file')
keywords = dict( [(line[5:-1],line[:5]) for line in f.readlines()] )
f.close()

keywords[my_word]

答案 5 :(得分:0)

很有可能,在具有未知长度记录的已排序文本文件上执行二进制搜索,通过重复二等分范围,并向前读取行终止符,效率略有下降。这是我在第一个字段中通过一个带有2个标题行的csv文件查找的内容。给它一个打开的文件,以及要查找的第一个字段。为您的问题修改它应该相当容易。偏移零点的第一行匹配将失败,因此可能需要特殊配置。在我的情况下,前两行是标题,并被跳过。

请原谅我下面没有抛光的蟒蛇。我使用此函数和类似函数直接从Maxmind分发的CSV文件中执行GeoCity Lite纬度和经度计算。

希望这有帮助

========================================

# See if the input loc is in file 
def look1(f,loc):
# Compute filesize of open file sent to us
hi = os.fstat(f.fileno()).st_size
lo=0
lookfor=int(loc)
# print "looking for: ",lookfor
while hi-lo > 1:
    # Find midpoint and seek to it
    loc = int((hi+lo)/2)
    # print " hi = ",hi," lo = ",lo
    # print "seek to: ",loc
    f.seek(loc)
    # Skip to beginning of line
    while f.read(1) != '\n':
        pass
    # Now skip past lines that are headers
    while 1:
        # read line
        line = f.readline()
        # print "read_line: ", line
        # Crude csv parsing, remove quotes, and split on ,
        row=line.replace('"',"")
        row=row.split(',')
        # Make sure 1st fields is numeric
        if row[0].isdigit():
            break
    s=int(row[0])
    if lookfor < s:
        # Split into lower half
        hi=loc
        continue
    if lookfor > s:
        # Split into higher half
        lo=loc
        continue
    return row  # Found
# If not found
return False