sed命令过滤特定的行，如何去掉前面的数字和横杠，或按set nu显示真实行数

lusklusklusk

文件alert_DW.log内容vi后如下
Sun Sep 18 10:22:15 2016
Sat Sep 17 10:07:46 2016
DBMS_STATS: GATHER_STATS_JOB
Errors in file
ORA-=10000: Unable to set values10000
Sat Sep 17 10:29:41 2016
ttt
fff
ddf
Errors in file
ORA-20001: Unable to set values20001
ddd
fff
ttt
mmm
nnn
fff

:set nu则显示如下
  1 Sun Sep 18 10:22:15 2016
  2 Sat Sep 17 10:07:46 2016
  3 DBMS_STATS: GATHER_STATS_JOB
  4 Errors in file
  5 ORA-=10000: Unable to set values10000
  6 Sat Sep 17 10:29:41 2016
  7 ttt
  8 fff
  9 ddf
10 Errors in file
11 ORA-20001: Unable to set values20001
12 ddd
13 fff
14 ttt
15 mmm
16 nnn
17 fff
18


执行命令sed -n "7,20 p" alert_DW.log|grep  -nA3 -nB2 'ORA-'
显示如下
3-ddf
4-Errors in file
5:ORA-20001: Unable to set values20001
6-ddd
7-fff
8-ttt


第一个问题：上面命令的如何把显示结果3-、4-、5：、6-、7-、8-剔除，显示为如下
ddf
Errors in file
ORA-20001: Unable to set values20001
ddd
fff
ttt


第二个问题，如何按:set nu显示真实行数，显示为如下
9 ddf
10 Errors in file
11 ORA-20001: Unable to set values20001
12 ddd
13 fff
14 ttt

lusklusklusk

第一个，找到答案了
grep后面的-nA中把n去掉就可以了

moperyblue

1. cat -n alert_DW.log|sed '6,$!d'|grep -B2 -A3 ORA-

复制代码