Shell Programming and Scripting

As I mentioned in my second comment, I cannot copy and paste the actual inputs in here as the forum complaints that there are too many icons.

I think, I might have confused with my modified output file. What I am after is when I print field 2, the second field you have mentioned should be printed. But unfortunately awk only prints portion of that SOAP XML as there are spaces and newlines. Below is what I get

Code:

<soapenv:Envelope xmlns:soapenv=""http://schemas.xmlsoap.org/soap/envelope/"" xmlns:cai3=""http://schemas.centra.com/cai3g1.2/"" xmlns:opt=""http://schemas.centra.com/""><soapenv:Header><cai3:SessionId>abc_4d728b7db4cf4950b660c37c0ccad7bd</cai3:SessionId><cai3:TransactionId>1605271108449435658</cai3:TransactionId></soapenv:Header><soapenv:Body>

Attached output.txt file is a manually edited file, where I want the entire field 2 with white space and newlines removed

You have confused us because every statement you have made has been contradicted by your examples. You have given us an input text file that is not a text file. You say that your input file is a CSV file (which is normally a set of <newline> delimited records containing comma separated fields). But your CSV file has comma separated records and no field separators or delimiters. And, you say you want to print the 2nd field, but you want to print the 2nd record. You say that you want to remove whitespace and newlines; but your sample output file contains multiple (sometimes even adjacent) space characters. (And, if all whitespace characters were removed, some of the tags in the XML data included in the second record in your "CSV" file would be changed invalidating your XML schema.)

According to the standards, the behavior of awk (and grep and sed) is undefined if the input files being processed are not text files, and (since neither you sample input nor your desired output file contain terminating <newline> characters) your sample input file is not a text file. But, since you seem to only want to process record 2 in a 3+ record input file, the following awk script may be able to get something reliably even though the input is not a text file since it will exit before it gets to the end of the input. As you requested, the output produced by this script is not a "line" and has no trailing (nor embedded) <newline> characters.

You said you want to remove whitespace characters from your output. Instead of doing that, the following converts each string of one or more whitespace characters in the input to a single <space> character in the output.

So, the following awk script does something sort of like what you have requested that might or might not be useful. If this doesn't work for you, PLEASE CLEARLY EXPLAIN EXACTLY WHAT YOU ARE TRYING TO DO and PROVIDE SAMPLES THAT MATCH THE EXPLANATION YOU PROVIDE (in CODE tags):

Code:

awk -v RS=',' '
NR == 2 {
	gsub(/[[:space:]]+/, " ")
	printf("%s", $0)
	exit
}' input.txt

If you want to try this code on a Solaris/SunOS system, change awk to /usr/xpg4/bin/awk. (Note that nawk won't work for this script on Solaris/SunOS systems either.)

Note that I named the input file input.txt because that is the name of the input file you supplied. But, I strongly object to naming any file that is not a text file with the filename extension .txt!

Thanks again for persisting with my confusing comments

You latest code works, but again because of my poor description the results are not as exactly what I expected when I have multiple records

To clear things out
1. Fields are separated by ,
2. Records are separated by " at start of a line

The sample input I provided had only 1 record and 3 fields