sed/awk group on first column

Eddie Adams · 04-09-2009, 07:53 AM

current:
15551256 user nl 1 Query SELECT SQL_CALC_FOUND_ROWS
15551256 user nl 2 Query SELECT SQL_CALC_FOUND_ROWS
15551256 user nl 3 Query SELECT SQL_CALC_FOUND_ROWS
15551256 user nl 4 Query SELECT SQL_CALC_FOUND_ROWS
15551330 user nl 0 Query SELECT SQL_CALC_FOUND_ROWS
15551367 user nl 0 Query SELECT SQL_CALC_FOUND_ROWS
15551372 user nl 0 Query SELECT SQL_CALC_FOUND_ROWS
15551486 user nl 1 Query SELECT SQL_CALC_FOUND_ROWS
15551486 user nl 2 Query SELECT SQL_CALC_FOUND_ROWS

desired:
15551256 user nl 4 Query SELECT SQL_CALC_FOUND_ROWS
15551330 user nl 0 Query SELECT SQL_CALC_FOUND_ROWS
15551367 user nl 0 Query SELECT SQL_CALC_FOUND_ROWS
15551372 user nl 0 Query SELECT SQL_CALC_FOUND_ROWS
15551486 user nl 2 Query SELECT SQL_CALC_FOUND_ROWS

jschiwal · 04-09-2009, 08:29 AM

What criteria are you using? Where there are duplicates, what determines which one to use?

If you simply want a line where the first column is unique, look at the uniq program. If you are testing for a range of values in the first column, then use awk.

Eddie Adams · 04-09-2009, 09:59 AM

The first column is unique. The fourth column has a max value.

colucix · 04-09-2009, 10:06 AM

Code:

sort -nr file | uniq -w8 | tac

Eddie Adams · 04-09-2009, 10:23 AM

Thanks.