UNIX for Beginners Questions & Answers

thank you rudic it worked. appreciate your help.

You can incorporate many-liners as well as one-liners into existing scripts:

Code:

...
var=$(cmd1 | sed '<one-line-here>' | cmd2)
...

or several lines like this. Notice the punctuation and that the indentation makes it easier to discern what belongs to where. The same goes for any text filter other than sed (awk, grep, ...) too:

Code:

...
var=$( \
       cmd1 |\
       sed '<one-line>
            <next-line>
            <etc.>' |\
       cmd2 \
)
...

I write programs in some form for 40 years now. I found that, as a rule of thumb, lines longer than 80 characters are usually wrong. Even if they aren't they are hard to read and should be avoided. Here is a real example from one of my scripts:

Code:

sed 's/[^\]#.*$//
     s/^#.*$//
     s/\\#/#/g
     /^[        ]*$/d
     s/[        ]*$//
     s/^/x/' "$fInFile" |\
while read chLine ; do
     chItem=""
     chValue=""
     ...

Readable, no? Here is the one-liner-version:

Code:

sed 's/[^\]#.*$//;s/^#.*$//;s/\\#/#/g;/^[        ]*$/d;s/[        ]*$//;s/^/x/' "$fInFile"|while read chLine;do;chItem="";chValue="";...

Probably not quite easy to decipher, is it?

I hope this helps.

bakunin

Note: in this case the newline escapes (\) can be omitted:

Code:

...
var=$( 
       cmd1 |
       sed '<one-line>
            <next-line>
            <etc.>' |
       cmd2
)
...