求简洁sed或者awk命令

liying_gg

1. 
   
2. 文件如下
   
3. C2*C4  as C1,
   
4. ,(a.OUTTRY_NUM+d.PP_REQ)/(e. CALLER_CE_SCC_NUM+a.INTRY_NUM-a.CALLEE_HLR_FAIL)*100 as C2
   
5. ,(a.LCALLEE_ANS_NUM+ c.LOC_BUSY_NUM+ c.LOC_NOANS+ c.LOC_REL_ALERT+ c.LOC_REJ_NUM+ c.LOC_AF_NUM)/(a.INTRY_NUM+ a.OUTTRY_NUM2+ a.OTHER_TRY+a.REDREQ_NUM+a.DL_REQ_NOANS- a.OUTTRY_NUM- a.HLR_FAIL_NUM- a.OUTTRY_NUM1- (c.early_rel+c.CALLER_STR_RES+c.ERCALL_NUM+c.LOC_OPER_SUCC))*100 as C3
   
6. ,(e. CALLER_CE_REQ_NUM/(e. CALLER_CE_REQ_NUM+e. CALLEE_CE_REQ_NUM)*100)*C7+(1-(e. CALLER_CE_REQ_NUM/(e. CALLER_CE_REQ_NUM+e. CALLEE_CE_REQ_NUM)*100))*C10*C8 as C4
   
7. ,(a.INCON_NUM+a.OUTCON_NUM)/(a.INTRY_NUM+a.OUTTRY_NUM)*100 as C5
   
8. ,(a.INANS_NUM+a.OUTANS_NUM)/(a.COME_TRY_NUM+a.OUTTRY_NUM)*100 as C6
   
9. ,e. CALLER_CE_SCC_NUM/e. CALLER_CE_REQ_NUM*100 as C7
   
10. ,C11.7/e. CALLEE_CE_REQ_NUM*100 as C8
    
11. ,e. CE_SCC_NUM_ALL/e. CE_REQ_NUM_ALL*100 as C9
    
12. ,d.MSC_P_RES/d.MSC_P_NUM*100 as C10
    
13. ,e. CE_SCC_NUM/e. CALL_TRY_NUM*100 as C11
    
14. ,a.PKG_FAIL_NUM/(a.PKG_NUM+ a.PKG_FAIL_NUM)*100 as C12
    
15. ,b.RTP_RPKG_LOST/(b.RTP_RPKG+b.RTP_RPKG_LOST)*100 as C13
    
16. ,0 as C14
    
17. ,0 as C15
    
18. 
    
19. 希望目标内容为grep -o -e "a\.[a-zA-Z_]*" file的结果，
    
20. 我用sed写了一个，sed -n '/a\.[a-zA-Z_]/!d;:a;h;s/\(.*\)\(a\.[a-zA-Z_]*\).*/\2/p;x;s/\(.*\)\(a\.[a-zA-Z_]*\).*/\1/;ta'
    
21. 写的太复杂了，谁有好点的逻辑？
    

复制代码

[ 本帖最后由 liying_gg 于 2009-12-22 17:06 编辑 ]

ly5066113

sed -nr ':a;s/([^\n]+)(a\.\w+)[^\n]*/\1\n\2/;ta;s/[^\n]*\n//p' urfile

btw: 楼主的grep有点小问题，a.OUTTRY_NUM2 后面的数字2应该带出来的。

yazi0127

你这儿是哪来的文件，呵呵。

liying_gg

好，比我的简洁，谢谢指正错误。

ly5066113

应该是某数据库存储过程中的一段，没有用到什么特别的函数，看不出是什么数据库。

liying_gg

一个哥们儿问我的问题，搞电信的。

curclew

还是用awk最方便
awk -v RS='a\\.\\w*' '{ print RT}' file

sed比较短的可用:
sed 's/a\.\w*/\n&\n/g;' file|sed '/^a\.\w*$/!d;'

liying_gg

有想法，不错。