Shell Programming and Scripting

You can use awk also :

Code:

$ cat file
something blah1.csv blah2.csv blah3.csv somethingelse
something blah10.csv blah20.csv blah300.csv somethingelse
something blah100.csv blah200.csv blah300.csv somethingelse
$ cat file | awk 'BEGIN{RS=" "}/.csv$/{printf("%s\n",$1)}'
blah1.csv
blah2.csv
blah3.csv
blah10.csv
blah20.csv
blah300.csv
blah100.csv
blah200.csv
blah300.csv

Code:

awk 'BEGIN{RS=" "}/.csv$/{printf("%s ",$1)}'

Basically awk has this syntax:
awk 'pattern action pattern action ....'

BEGIN and END are special patterns to get awk to execute it's corresponding actions before and after the data stream.

In our case, the BEGIN block instructed awk to treat the record separator to be a single space. The default is newline.

For every record (remember your RS is " "), the pattern to match is /.csv$/ - the record has to be ended with .csv. The dollar sign ($) is an anchor to indicate end of line. If it matches the pattern, print the first field. Since you want it to be in a single line, I use printf instead of print so that it does not produce a newline.

Hope the explanation is clear.

AWK is powerful. My blog
Chi Hung Chan
featured a lot of real life examples of using AWK.
Here is the search link URL:
Chi Hung Chan