怎么去掉重复的时间

yinyuemi

回复 8# bblxl2012

1. sed -rn 'h;:a;n;/^\|/H;/^[^|]/{x;bb};${x;bb};ba;:b;s/([^|]+\|([^|]+)\|[^\n]+)(.*)\n\|\2(\|.*)/\1\3\n|\4/;tb;p;ba'

复制代码

bblxl2012

yinyuemi 发表于 2013-07-12 10:51
回复 8# bblxl2012

谢谢。辛苦了。能解释一下吗？

yinyuemi

回复 12# bblxl2012

1. sed -rn '
   
2. h;  # store 1st line to hold space
   
3. :a; # set label "a"
   
4.    n; # read the next line
   
5.    /^\|/H; # if the line starts with '|', append the current line to hold space
   
6.    /^[^|]/{x;bb}; # if line starts without '|' or reading last line (the next command line), do command "x" to exchange the context in hold space and pattern space, then branch to label "b";
   
7.    ${x;bb};
   
8. ba; # branch to label "a"
   
9. :b; # set label "b";
   
10.    s/([^|]+\|([^|]+)\|[^\n]+)(.*)\n\|\2(\|.*)/\1\3\n|\4/; # treat the context in the pattern space by s/// to fit your aim
    
11. tb; # if s/// has done successfully, branch to label "b"; otherwise, execute the next command line;
    
12. p; # print the context in the pattern space
    
13. ba # branch to label "a"
    
14. '

复制代码