sed/awk/perl: find a regex, copy 5 columns of this

问题:

I have following lines:

057         -   -   No    adod3 stptazlqn    10    753
tlm    10    027        
stp    10    021                
12         -   -   No    azad1
bbcz     30    12        
03085         -   -   No    azad1 azad1222
xxaz    1    12        
azzst    1    12        
hss     2    12 

what I need to do is:

1. Find lines starting with a number [0-9].
2. Copy the first 5 columns separated by a space ' '.
3. Paste it in the next lines not starting with a number.

057         -   -   No    adod3 stptazlqn    10    753
057         -   -   No    adod3     tlm    10    027        
057         -   -   No    adod3     stp    10    021                
12         -   -   No    azad1
12         -   -   No    azad1    bbcz     30    12        
03085         -   -   No    azad1 azad1222
03085         -   -   No    azad1    xxaz    1    12        
03085         -   -   No    azad1    azzst    1    12        
03085         -   -   No    azad1    hss     2    12

Thanks for any help in advance.

Cheers, Slaw

回答1:

awk to the rescue!

$ awk '{if($1+0==$1) p=$1 FS $2 FS $3 FS $4 FS $5; else $0=p FS $0}1' file | column -t

057    -  -  No  adod3  stptazlqn  10  753
057    -  -  No  adod3  tlm        10  027
057    -  -  No  adod3  stp        10  021
12     -  -  No  azad1
12     -  -  No  azad1  bbcz       30  12
03085  -  -  No  azad1  azad1222
03085  -  -  No  azad1  xxaz       1   12
03085  -  -  No  azad1  azzst      1   12
03085  -  -  No  azad1  hss        2   12

here FS is the field separator by default is the white space. Number test is done with zero addition; pipe to column for pretty formatting. The 1 at the end is the shorthand for printing the line ($0 variable)

回答2:

This might work for you (GNU sed):

sed -r '/^[0-9]/!G;s/(.*\S)\s*\n((\S+\s*){5}).*/\2 \1/;h' file

Use pattern matching to extract the fields from line beginning with an integer which will have been appended to a line that did not.

To align the columns correctly use:

sed -r '/^[0-9]/!G;s/(.*\S)\s*\n((\S+\s*){5}).*/\2 \1/;h' file | column -t