CODEWITHSUNDEEP
linuxshellterminalnewlineoutput
Edd
Edd

Reputation: 685

(BASH)How to remove default 'newline' operator in command

I have a command:

find /etc -type f \( ! -perm /o=w \) -exec /usr/bin/ls -hastl '{}' \; -exec /usr/bin/md5sum '{}' \; > somefile.log

and this command creates such printout (i need to have no breaking line) : 

...
4.0K -rw-r--r--. 1 root root 332 Oct 23  2012 /etc/xinetd.d/rsync
fd32314f3157aaf15712e6da7758060d  /etc/xinetd.d/rsync
....

question: what I must add to this command to remove new line operation to get such result:

(...) is '-rw-r--r--. 1 root root 332 Oct 23  2012'

...
4.0K (...) /etc/xinetd.d/rsync >> fd32314f3157aaf15712e6da7758060d  /etc/xinetd.d/rsync
...

Upvotes: 1

Views: 949

Answers (3)

mschilli
mschilli

Reputation: 1894

Two join every pair of adjacent lines just pipe your output to awk 'NR%2{line=$0;next}{print line sep $0}' sep=" >> " adjusting sep to your need.

Addendum:

In response to the original author's comment I am going to explain some details of this solution:

awk processes input record-wise, without further options that is line-by-line. The (internal) variable NR stores the current record's number. % is the modulus operator, so NR%2 evaluates to 1 for odd line numbers and 0 for even line numbers. {...} groups commands (so called 'action') to perform. Since 1 is considered "true" and 0 "false" the first action is only performed for odd line numbers. This action consists of two commands: First, the whole line ($0) is stored in the variable line and then the next record is processed (without considering any other action). For even lines, this action is not performed, so the second action is considered. Since there is no expression before the { the action is performed: the stored (i.e. previous) line is concatenated with a variable called sep and the current line ($0). The resulting string is printed (and automatically appended by one new-line in the end). The value of sep is handed to awk as a parameter to make it easy to adjust it without messing with the script. Since awk reads standard input and output without further arguments you can just pipe (|) the output of your find to awk (find ... -exec ... -exec ... | awk ... > ...)

Upvotes: 1

marderh
marderh

Reputation: 1256

IMHO You can only execute one command with exec. A possible solution might be you execute "sh -c" so you can use pipes and similar:

With one newline at the end:

find /etc -type f -exec sh -c "/bin/ls -hastl '{}' | tr '\n' ' '" \; -exec /usr/bin/md5sum '{}' \;

without any newline:

find /etc -type f -exec sh -c "/bin/ls -hastl '{}' | tr '\n' ' '" \; -exec sh -c "/usr/bin/md5sum '{}'| tr -d '\n' " \;

Upvotes: 0

Some programmer dude
Some programmer dude

Reputation: 409136

How about a command such as

echo "$(ls -hastl '{}') >> $(md5sum '{}')"

Upvotes: 1

Related Questions