help on awk |merge two words

mad_penguin · 04-08-2011, 08:01 AM

Heloo

I have a file that contains name like, firs_name last_name, on a row (

ex: John Smith
Michael Jones
Maria Lee

I need something to merge firs_name + last_name by putting a '.' (dot) between, like: John.Smith etc.

I think awk can do it but I don't know how.

Any help would be much appreciated.

Thanks

smallpond · 04-08-2011, 08:13 AM

if there are always exactly two names separated by a space, awk is fine:

Code:

 cat sample
John Smith
Michael Jones
Maria Lee

 awk  'BEGIN {OFS="."} {print $1,$2}' < sample 
John.Smith
Michael.Jones
Maria.Lee

This changes the output field separator to ".", then prints the first two fields of each input line to each output line. The output record separator (newline) is printed at the end of each line.

colucix · 04-08-2011, 08:19 AM

For any number of fields:

Code:

awk -v OFS=. '$1=$1'

the -v part is the same as the BEGIN section in the previous example. The $1=$1 expression forces awk to split the record into fields, so that the default action (print) actually uses the specified output field separator.

mad_penguin · 04-08-2011, 08:21 AM

Thanks a lot mate!

MTK358 · 04-08-2011, 09:14 AM

Mark the thread as solved.

By the way, you can also do it with sed:

Code:

sed 's/[^ ]\+ [^ ]\+/\1.\2/'

wpeckham · 04-08-2011, 09:50 AM

Actually, awk (while it will work) may not be the best tool for the job. Multiple solutions come to mind involving perl, tr, and simple bash scripting.

PMorph · 04-09-2011, 08:31 AM

Care to elaborate? colucix's solution is hardly too much of a trouble to implement.

grail · 04-09-2011, 08:47 AM

Quote:

Care to elaborate? colucix's solution is hardly too much of a trouble to implement.

I agree with PMorph. The only extra advantage you might have of using say sed is not having to use a temp file assuming you are making
alterations to the original:

Code:

sed -i 's/ /./' file

ghostdog74 · 04-09-2011, 10:41 AM

Code:

$ awk '$1=$1' OFS="." file

ghostdog74 · 04-09-2011, 10:42 AM

Quote:

Originally Posted by wpeckham

Actually, awk (while it will work) may not be the best tool for the job. Multiple solutions come to mind involving perl, tr, and simple bash scripting.

Nonsense. awk is there to reduce excessive bash scripting. Perl is over-rated when it comes to file parsing.

ghostdog74 · 04-09-2011, 10:43 AM

Quote:

Originally Posted by grail

I agree with PMorph. The only extra advantage you might have of using say sed is not having to use a temp file assuming you are making
alterations to the original:

Code:

sed -i 's/ /./' file

where is the extra advantage?

Code:

$ awk  '{$1=$1; print > FILENAME}' OFS="." file