怎么可以去掉这些乱七八糟的字符？

forest077

我想了一招，可以解决楼上那位同学说的两种情况，在sco下面已经经过测试。请楼主和你的同学们试试。

1. 
   
2. sed -n '/Status/,/\/TD/p' file|sed 's/\(<TD>\)/\1\
   
3. /;s/\(<\/TD>\)/\
   
4. \1/'|sed '/^[[:space:]]*$/d'|sed -n '/<TD>/{n;p;}'
   

复制代码

Y_

只能用两个字来形容 - 厉害!

Kevinzj

这么长的script, 厉害啊. 谢谢.
但是中间处有问题, 说 sed: command garbled: s/\(<TD>\/\1\/;s/\(<\/TD>\)/\\1/
我猜forest077的意思是要除掉两头的<TD></TD>, 所以用回了我的老code  sed s'/<TD>\(.*\)<\/TD>/\1/', 不知是否符合forest077的原意.
而我这样做也没有得出结果, 连 CLOSED 都过滤掉了. 问题应该是在最后一个sed里,
我做到sed -n '/>Status:/,/\/TD/p' "show_bug.cgi?id=9998" |sed s'/<TD>\(.*\)<\/TD>/\1/'|sed '/^[[]]*$/d'的时候结果如下:

<TH><B><A HREF="/scdocs/issue_lifecycle.html" onclick="return launch(this.href, 1)" title="Note: link may open in new window" class="helplink">Status:</A></B></TH>
        CLOSED

但加上sed =n '/<TD>/{n;p;}' 则无所得, 因为已经没<TD>了.
我固然可以用 sed '/Status/d'之类的方法去掉这一行来得出结果, 但我再想想觉得这方法应该不是forest077的原意, 能否请你说明一下你的code呢, 谢谢了.

Kevinzj

喔, 我想我明白了, 先是找出从Status到/TD之间的行, 然后消掉<TD></TD>, 再去空行, 然后再把<TD>的下一行弄出来. 对吧? 果然是高手啊, pf pf

forest077

需要注意的是，我的代码是三行，不能并到一行去，三行之间的回车换行是必须的，是作为文本替换用的，不是为了好看的。
思路是先找出status到/TD之间的行，然后在TD后面换行，在/TD之前换行，然后消掉空行，然后找到TD的下面一行。经过一系列处理，这样就可以处理原文件中四个要素之间位置的任意关系了。

admirer

原帖由 "maxx"][code 发表：
10018 CLOSED FIXED
1018 CLOSED FIXED
1019 CLOSED FIXED
1013 CLOSED WONTFIX Closing as WONTFIX.</PRE>
10133 CLOSED FIXED
10134 CLOSED FIXED
10135 CLOSED FIXED Closing as FIXED.</PRE>
10136 CLOSED FIXED
10138 CLOSED FIXED
10139 CLOSED FIXED
10141 CLOSED FIXED
1014 CLOSED FIXED
10143 CLOSED FIXED
10144 CLOSED FIXED
10145 CLOSED FIXED
.
.
[/code]

1. sed -n 's/^\([0-9]\{1,\}\) \(\([A-Z]\{1,\} \{0,1\}\)\{1,2\}\).*/\1 \2/p' filename
   
2. .
   
3. .
   
4. .
   
5. 197 RESOLVED FIXED
   
6. 1973 RESOLVED FIXED
   
7. 1975 RESOLVED FIXED
   
8. 1979 CLOSED FIXED
   
9. 1985 CLOSED FIXED
   
10. 13007 CLOSED FIXED
    
11. 13047 CLOSED WORKSFORME
    
12. 13057 CLOSED INVALID
    
13. 13058 CLOSED DUPLICATE
    
14. 13085 CLOSED FIXED
    
15. 13090 CLOSED FIXED
    
16. 13097 CLOSED FIXED
    
17. 13107 CLOSED FIXED
    
18. 13134 NEW
    
19. 13169 CLOSED FIXED
    
20. 133 RESOLVED WONTFIX
    
21. 138 CLOSED INVALID
    
22. 1348 NEW
    
23. 1356 CLOSED FIXED
    
24. 1369 CLOSED DUPLICATE
    
25. 138 CLOSED WORKSFORME
    
26. 13365 CLOSED WONTFIX
    

复制代码