Shell Programming and Scripting

try sed+awk:

Code:

sed -i "1s/.*/&`awk 'BEGIN {clientlist = "";FS=":"} /client/&&NR>1 { clientlist = clientlist $2 }END { print clientlist"]"}' file`/" file

Thanks again. That worked from the command line when I added some empty single quotes after the -i. However, I'd like it to be executable from within a bash script, as it will be one stage in a longer conversion process. With sed scripts, I can reference the filename (passed in from cmd line) as $1, but I don't think that works with awk, as the $n syntax references fields. Is there an alternate way of referring to the filename?

I have some other questions about the syntax, for my own understanding:

Code:

"1s/.*/&`awk '...'

...I assume here the & is acting as a concatenate operator, so this is saying "add the results of the following awk script to line 1 (1s) after all the existing characters (.*)" ?

Code:

/client/&&NR>1

match when text 'client' appears AND line number > 1 ?

Code:

`awk '...' file`/

...here my question is whether the ` quotes are the standard notation for delimiting an awk script that is feeding a sed script, or are they used because ' and " are already used elsewhere in the script? Or is it like with regex, where any character could be used to delimit, so long as it matches at both ends?

Also, what does the / at the end denote?

Thanks again for all your help, I am learning alot

You can also have a look at this link

Found it, I just needed to add

filename="$1"

at the top of my bash script, then refer to $filename thereafter.

thanks again yinyuemi, and thanks for the link ctsgnb - that helped with some substr/length stuff I needed to do.

And this one is just in case you want to play with sed

OK, the saga continues. I am nearly done, but have this structure left, e.g.:

Code:

"one":"field one",
"two":"field two",
"three":"
field three line 1

field three line 2

field three line 3
"

and I need to replace the line breaks in that last field with text '\n'. However when I use awk to access $s of line '/three/', I just get '"'.

I don't want to just replace all line breaks with '\n' in the file, because the ones at the end of lines one and two are not harming the input, whereas the ones in field three are (I'm trying to json_decode in PHP).

How would I go about achieving this?

how about this?

Code:

awk '/\"/&&$0!~/,/{++p}{printf (p%2?$0" ":$0"\n")}' file
 
"one":"field one",
"two":"field two",
"three":"field three line 1 field three line 2 field three line 3"