25

A string maybe this 

ipath= "./data/NCDC/上海/虹桥/9705626661750dat.txt"

or this

ipath = './data/NCDC/ciampino/6240476818161dat.txt'

How do I know the first string contains chinese?

I find this answer maybe helpful: Find all Chinese text in a string using Python and Regex

but it didn't work out: 

import re
ipath= "./data/NCDC/上海/虹桥/9705626661750dat.txt"
re.findall(ur'[\u4e00-\u9fff]+', ipath) # => []
Community
  • 1
  • 1
ZK Zhao
  • 19,885
  • 47
  • 132
  • 206

8 Answers8

29

The matched string should be unicode as well

>>> import re
>>> ipath= u"./data/NCDC/上海/虹桥/9705626661750dat.txt"
>>> re.findall(r'[\u4e00-\u9fff]+', ipath)
[u'\u4e0a\u6d77', u'\u8679\u6865']
xecgr
  • 5,095
  • 3
  • 19
  • 28
  • 1
    Some reason, add `u` gives me a syntax error, remove it works for me `re.findall(r"[\u4e00-\u9fff]+", ipath)` – LYu May 13 '18 at 03:50
  • @xecgr I get the same problem (and solution) mentioned by @LYu above. Please can you explain what the `u` is required for and why this is causing problems that can be resolved by removing it from your code? – Matt Shepherd Nov 29 '18 at 17:07
  • Sorry, what does the reason? How do you get u4e00? – Alston Dec 18 '22 at 12:04
12

If you just want to know whether there is a chinese character in your string you don't need re.findall, use re.search and the fact that match objects are truthy.

>>> import re
>>> ipath= u'./data/NCDC/上海/虹桥/9705626661750dat.txt'
>>> ipath2 = u'./data/NCDC/ciampino/6240476818161dat.txt'
>>> for x in (ipath, ipath2):
...     if re.search(u'[\u4e00-\u9fff]', x):
...         print 'found chinese character in ' + x
... 
found chinese character in ./data/NCDC/上海/虹桥/9705626661750dat.txt
timgeb
  • 76,762
  • 20
  • 123
  • 145
9

Use \p{Han} regex with PyPi regex:

import regex
ipath = "./data/NCDC/上海/虹桥/9705626661750dat.txt"
print(regex.findall(r'\p{Han}+', ipath) )
# => ['上海', '虹桥']

See Python proof.

The regex.search is enough to detect:

if regex.search(r'\p{Han}', ipath):
    print(f'"{ipath}" contains Chinese!')
Ryszard Czech
  • 18,032
  • 4
  • 24
  • 37
  • 1
    Came here to add the same answer. The `regex` module is often needed anyway, and this way you don't need to keep double-checking that you've got the write character ranges in your code. Plus if a new block of Han characters are added in the future, you're likely to get your code updated via a dependency update automatically. – Nate Glenn Nov 21 '22 at 12:36
6

And for those of us who don't care for re:

>>> ipath= u"./data/NCDC/上海/虹桥/6240476818161dat.txt"
>>> for i in range(len(ipath)):
...  if ipath[i] > u'\u4e00' and ipath[i] < u'\u9fff':
...   print ipath[i]
... 
上
海
虹
桥

Edit: for the full list of Chinese characters this SO link is worth looking at as the range U+4E00..U+9FFF is not complete. What's the complete range for Chinese characters in Unicode?

brandon
  • 43
  • 5
Rolf of Saxony
  • 21,661
  • 5
  • 39
  • 60
  • This is simply taking what was done at the C level and moving it to Python. For large blocks of text, it will be significantly slower. – Brad Solomon Jul 29 '19 at 16:34
  • 1
    @BradSolomon and your point is? This was a non `re` response, it doesn't claim to be anything else, certainly not faster. – Rolf of Saxony Jul 29 '19 at 16:50
4

Using these codepoint ranges, we can write an is_cjk function:

# list of cjk codepoint ranges
# tuples indicate the bottom and top of the range, inclusive
cjk_ranges = [
        ( 0x4E00,  0x62FF),
        ( 0x6300,  0x77FF),
        ( 0x7800,  0x8CFF),
        ( 0x8D00,  0x9FCC),
        ( 0x3400,  0x4DB5),
        (0x20000, 0x215FF),
        (0x21600, 0x230FF),
        (0x23100, 0x245FF),
        (0x24600, 0x260FF),
        (0x26100, 0x275FF),
        (0x27600, 0x290FF),
        (0x29100, 0x2A6DF),
        (0x2A700, 0x2B734),
        (0x2B740, 0x2B81D),
        (0x2B820, 0x2CEAF),
        (0x2CEB0, 0x2EBEF),
        (0x2F800, 0x2FA1F)
    ]

def is_cjk(char):
    char = ord(char)
    for bottom, top in cjk_ranges:
        if char >= bottom and char <= top:
            return True
    return False

Which we can then use to process text, using functions like filter, any, all, and map to process the text character-by-character, or compose more complex functions:

txt = "./data/NCDC/上海/虹桥/9705626661750dat.txt"
txt_sanitized = "./data/NCDC/9705626661750dat.txt"
any(map(is_cjk, txt)) # True
any(map(is_cjk, txt_sanitized)) # False
''.join(filter(is_cjk, txt)) # '上海虹桥'

Note that the CJK ranges will include not only Chinese characters but also may include Korean and Japanese characters. For more complex usage, try a dedicated library like cjklib.

9999years
  • 1,561
  • 13
  • 14
2
import re
ipath= raw_input()
print re.findall(ur'[\u4e00-\u9fff]+', ipath.decode("utf-8"))

Output:./data/NCDC/上海/虹桥/9705626661750dat.txt [u'\u4e0a\u6d77', u'\u8679\u6865']

You need to decode the input to make it unicode.

or

 import re
 ipath= unicode(raw_input(),encoding="utf-8")
 print re.findall(ur'[\u4e00-\u9fff]+', ipath)
vks
  • 67,027
  • 10
  • 91
  • 124
1

'' is a bytestring on Python 2. Either add from __future__ import unicode_literals at the top of the module or use unicode literals: u'':

>>> import re
>>> ipath= u"./data/NCDC/上海/虹桥/9705626661750dat.txt"
>>> re.findall(ur'[\u4e00-\u9fff]+', ipath)
[u'\u4e0a\u6d77', u'\u8679\u6865']
jfs
  • 399,953
  • 195
  • 994
  • 1,670
1

According to this question, the range should be [\u2E80-\u2FD5\u3190-\u319f\u3400-\u4DBF\u4E00-\u9FCC]

Kevin He
  • 1,210
  • 8
  • 19