怎么去掉多余的行？请帮忙

maxx

请教：如何去掉下面文件1 多余的文字，转换成 文件2的形式呢？

文件1：

1. 18776
   
2. CLOSED
   
3. 18783
   
4. CLOSED
   
5. 1881
   
6. STARTED
   
7. 18817
   
8. RESOLVED
   
9. 18835
   
10. REOPENED
    
11. in the state UNCONFIRMED, is there a method to this maddness?
    
12. Regarding the UNCONFIRMED state of this (and other) issues: Normally
    
13. But why set to RESOLVED INVALID?
    
14. 18843
    
15. CLOSED
    
16. 21385
    
17. CLOSED
    
18. Ho do you produce this NEW
    
19. NEW FIXED seems to result from reassigning fixed issues.
    
20. 21427
    
21. RESOLVED
    
22. 21441
    
23. RESOLVED
    
24. 21677
    
25. CLOSED
    
26. 21678
    
27. REOPENED
    
28. as RESOLVED/FIXED, because, due to the presence of
    
29. 21906
    
30. CLOSED

复制代码

文件2：

1. 18776
   
2. CLOSED
   
3. 18783
   
4. CLOSED
   
5. 1881
   
6. STARTED
   
7. 18817
   
8. RESOLVED
   
9. 18835
   
10. REOPENED
    
11. 18843
    
12. CLOSED
    
13. 21385
    
14. CLOSED
    
15. 21427
    
16. RESOLVED
    
17. 21441
    
18. RESOLVED
    
19. 21677
    
20. CLOSED
    
21. 21678
    
22. REOPENED
    
23. 21906
    
24. CLOSED

复制代码

多谢多谢

john_student

try :

grep -E '^[0-9]\{1,\}$|^[A-Z]*ED$' ur_file

bjgirl

1. $grep -A1 [0-9] a|sed '/--/d'
   
2. 18776
   
3. CLOSED
   
4. 18783
   
5. CLOSED
   
6. 1881
   
7. STARTED
   
8. 18817
   
9. RESOLVED
   
10. 18835
    
11. REOPENED
    
12. 18843
    
13. CLOSED
    
14. 21385
    
15. CLOSED
    
16. 21427
    
17. RESOLVED
    
18. 21441
    
19. RESOLVED
    
20. 21677
    
21. CLOSED
    
22. 21678
    
23. REOPENED
    
24. 21906
    
25. CLOSED
    

复制代码

bjgirl

[quote]原帖由 "john_student"]' ur_file[/quote 发表：

your script is not work
try :

1. awk '/^[0-9]/||/^([A-Z])+ED/' file

复制代码

john_student

对grep的匹配表达式真的搞不懂，
特别对^,$,{,},|的使用，还有加不加\的区别。

另外也没留意到每行末尾的空格。

改用sed了：

1. 
   
2. sed -n '/^[0-9]\+[ ]*$\|^[A-Z]\+ED[ ]*$/p'  ur_file
   

复制代码

bjgirl

[quote]原帖由 "john_student"][/quote 发表：

这个咋样? 一个while搞定

1. while read i;do ((${#i}<=8))&&echo $i;done<file
2. 18776
3. CLOSED
4. 18783
5. CLOSED
6. 1881
7. STARTED
8. 18817
9. RESOLVED
10. 18835
11. REOPENED
12. 18843
13. CLOSED
14. 21385
15. CLOSED
16. 21427
17. RESOLVED
18. 21441
19. RESOLVED
20. 21677
21. CLOSED
22. 21678
23. REOPENED
24. 21906
25. CLOSED

复制代码

btw:纯属即兴

maxx

长见识！

c1l2d3

这个很牵强:

1. sed 's/ *//' file|grep -v ' '

复制代码

forest077

楼上诸位，bjgirl的我没有看懂，john的判断ED不见得是个好办法。我想楼主是想要数字串紧接的一行。用sed有更好的办法的：

1. 
   
2. sed -n '/^[0-9]\{1,\}$/{N;p;}' aaa
   

复制代码

aaa是原数据文件。
还有，

原帖由 "bjgirl" 发表：

your script is not work

语法错误，应该是"does not work"吧，呵呵。

bjgirl

原帖由 "forest077" 发表：

语法错误，应该是"does not work"吧，呵呵。

我的那个while的,是根据$i的长度来判断,有些牵强