處理json中的內(nèi)容

hope1234567hope

將csv的轉(zhuǎn)換為json后，此時(shí)為tst.json文件如下：
[{"a":"10.100.121.187","b":"liux18","c":"2015-11-02 00:00:36","d":"GET /1jsp"},
{"a":"10.100.91.88","b":"liyunguo","c":"2012-01-01 00:01:19","d":"GET /p2.jsp"},
{"a":"10.208.26.203","b":"xuaiping1","c":"2010-12-01 00:01:31","d":"GET /in.jsp"},
{"a":"10.100.91.88","b":"liyunguo","c":"2013-12-31 00:01:32","d":"GET /n.jsp"},
{"a":"10.100.59.3","b":"liushubin2","c":"2014-12-14 00:02:46","d":"GET /2"},
{"a":"10.100.59.3","b":"liushubin2","c":"2011-12-12 00:11:46","d":"GET /1"}]

想將文件中所有的c中的內(nèi)容改為類似如下的（此舉例第一行的）
"c":"2015-11-02 00:00:36"  改為 "c" : {"$date" : "2015-11-02T00:00:36Z"}
請(qǐng)問(wèn)這個(gè)shell怎么寫？

elu_ligao

1. sed 's/\("c":\)\("[^ ]\+\) \([^"]\+\)"/\1{"$date":\2T\3Z"}/'
   

復(fù)制代碼

moperyblue

1. sed -r '/"c"/s/(.*"c":)(\S+)\s([^"]*)"/\1 {"$date" : \2T\3Z"}/'

復(fù)制代碼

reb00t

1. n [7]: d = {"a":"10.100.121.187","b":"liux18","c":"2015-11-02 00:00:36","d":"GET /1jsp"}
   
2. 
   
3. In [8]: d
   
4. Out[8]:
   
5. {'a': '10.100.121.187',
   
6. 'b': 'liux18',
   
7. 'c': '2015-11-02 00:00:36',
   
8. 'd': 'GET /1jsp'}
   
9. 
   
10. In [17]: d['c']= "%sT%sZ" % tuple(d['c'].split())
    
11. 
    
12. In [18]: d['c']
    
13. Out[18]: '2015-11-02T00:00:36Z'
    
14. 
    
15. In [19]: d['c']={'$date':d['c']}
    
16. 
    
17. In [20]: d
    
18. Out[20]:
    
19. {'a': '10.100.121.187',
    
20. 'b': 'liux18',
    
21. 'c': {'$date': '2015-11-02T00:00:36Z'},
    
22. 'd': 'GET /1jsp'}

復(fù)制代碼

python的代碼本身就能實(shí)現(xiàn)，python里面在加shell看起來(lái)不倫不類

hope1234567hope

正打算學(xué)習(xí)python，能寫一個(gè)Python的腳本么？我學(xué)習(xí)下，謝謝！回復(fù) 4# reb00t

jason680

回復(fù) 1# hope1234567hope

1. $ perl -pe '{s/("c":)"([^"]+)"/$1\{"\$date":"$2Z"\}/g}' tst.json
   
2. [{"a":"10.100.121.187","b":"liux18","c":{"$date":"2015-11-02 00:00:36Z"},"d":"GET /1jsp"},
   
3. {"a":"10.100.91.88","b":"liyunguo","c":{"$date":"2012-01-01 00:01:19Z"},"d":"GET /p2.jsp"},
   
4. {"a":"10.208.26.203","b":"xuaiping1","c":{"$date":"2010-12-01 00:01:31Z"},"d":"GET /in.jsp"},
   
5. {"a":"10.100.91.88","b":"liyunguo","c":{"$date":"2013-12-31 00:01:32Z"},"d":"GET /n.jsp"},
   
6. {"a":"10.100.59.3","b":"liushubin2","c":{"$date":"2014-12-14 00:02:46Z"},"d":"GET /2"},
   
7. {"a":"10.100.59.3","b":"liushubin2","c":{"$date":"2011-12-12 00:11:46Z"},"d":"GET /1"}]

復(fù)制代碼

hope1234567hope

發(fā)覺(jué)不行，主要是"c": "2015-11-02 00:00:36"的 "2015(也就是與：之間有個(gè)空格)前面有個(gè)空格，這樣此語(yǔ)句就起作用了，該怎么改sed呢？對(duì)sed確實(shí)不是很明白，謝謝回復(fù) 2# elu_ligao

elu_ligao

那就加個(gè)空格呀

1. sed 's/\("c":\s*\)\("[^ ]\+\) \([^"]\+\)"/\1{"$date":\2T\3Z"}/'

復(fù)制代碼

hope1234567hope

謝謝。能解釋下\("[^ ]\+\) \([^"]\+\)"/\1{"$date":\2T\3Z"}的意思么 ，謝謝了  ，確實(shí)對(duì)這個(gè)不是明白，見笑了
主要是\ 和 ）回復(fù) 8# elu_ligao

expert1

建議你了解下jq,這個(gè)才是處理json的標(biāo)準(zhǔn)。