Hi,
My output data from software tool is like
1234 84 test1 file:about user_detials(bankUser) :jpeg 10 40 xys
1356 2 branches dir:list of files(dirlisting:1) :directory 20 80 abc
and want this to be printed as below. where first and last 3 columns along with the string which is in the '()' brackets only needed.
1234 84 test1 bankUser 10 40 xys
1356 2 branches dirlisting:1 20 80 abc
------ Post updated at 11:45 AM ------
Also the column for string between brackets() are dynamic
Hello,
My initial thought was that this would be straightforward, and something like this one line of awk would work:
$ cat test.txt
1234 84 test1 file:about user_detials(bankUser) :jpeg 10 40 xys
1356 2 branches dir:list of files(dirlisting:1) :directory 20 80 abc
$ awk '{ gsub("\\(", " ") ; gsub("\\)", " ") ; print $1,$2,$3,$6,$8,$9,$10}' test.txt
1234 84 test1 bankUser 10 40 xys
1356 2 branches files :directory 20 80
$
So the idea is that we use awk , first replacing the brackets with spaces, and then just do a straightforward print of the fields we need.
But as you'll have seen, the output we get doesn't quite match what you want. It turns out that the reason for this is that the number of fields is inconsistent, and changes from one line of your test data to the other.
To be specific, the first line has nine fields, whereas the second line has ten:
$ head -1 test.txt | awk '{print NF}'
9
$ tail -1 test.txt | awk '{print NF}'
10
Now others with greater familiarity with awk than myself may well be able to still come up with an elegant one-line way to deal with this, but personally I would think that you have to make sure that the only spaces in your data are used strictly as field separators, or alternatively that any fields using spaces are at least quoted.
The following script would seem to do what you need:
$ cat script.sh
#!/bin/bash
file=test.txt
while read -r line
do
type1=`echo "$line" | awk -F\( '{print $2}' | awk -F\) '{print $1}'`
num1=`echo "$line" | awk -F\) '{print $2}' | awk '{print $2}'`
num2=`echo "$line" | awk -F\) '{print $2}' | awk '{print $3}'`
text1=`echo "$line" | awk -F\) '{print $2}' | awk '{print $4}'`
echo $line | awk -v type1=$type1 -v num1=$num1 -v num2=$num2 -v text1=$text1 '{print $1,$2,$3,type1,num1,num2,text1}'
done < "$file"
$ ./script.sh
1234 84 test1 bankUser 10 40 xys
1356 2 branches dirlisting:1 20 80 abc
$
So we use what field separators we can always count on (the brackets, or so I'm hoping !) to strip out the fields we need one at a time, then we build up our awk line at the end.
Hope this helps. If not, then if you can let me know in what respects it fails we can take things from there. The main thing in general with this type of problem is to make your input as consistently- and reliably-formatted as you possibly can, always.
What operating system are you using?
What shell are you using?
What have you tried to solve this problem on your own?
What is a "stripped one column string"?
Is it crucial that the varying numbers of spaces between the various columns in the input be duplicated in the output?
What, exactly, does "the column for string between brackets() are dynamic" mean? Will there ever be more than one pair of parentheses on an input line?
What operating system are you using?
linux(centos6)
What shell are you using?
bash/sh
What have you tried to solve this problem on your own?
still in early stages of awk so nothing I got on this
What is a "stripped one column string"?
Sorry if that was confusing but what I was trying to say it to print the line with few columns which are consistent(first and last 3 columns) and along with that in between data to be cut/stripped and print data that is lying between () brackets.
Is it crucial that the varying numbers of spaces between the various columns in the input be duplicated in the output?
Yes, as I said above, I need first and last 3 columns and the string mentioned inside () only as rest all doesn't make any sense to me and it sometimes goes too lengthy..
What, exactly, does "the column for string between brackets() are dynamic" mean? Will there ever be more than one pair of parentheses on an input line?
using the dynamic means the number of columns between first and last 3 columns are unpredictable as it's kind of description for developers but meaningless long data to me.
It'll be helpful If I can have single line command so that I can pipe to the output and get the desired results.
Again sorry for the confusion if any.
I fail to understand how changing varying numbers of <space>s between fields in your input lines (such as 1 <space> between most fields, but 2 <spaces> between the 1st and 2nd fields in the 2nd input line in your file) into single <space>s between fields would make it hard for you to identify the first three fields or the last three fields in the output you're trying to produce. This is especially confusing since the output that you say is needed for the 2nd line from your input file must create output that has 2 <space>s before the last three fields in the output when there was only 1 <space> in that position in your sample input file??? Insisting on this requirement does invalidate everything that drysdalk suggested. But, I don't understand how requiring additional <space>s in the output helps keep the output from being "too lengthy".
Please clearly explain why changing 2 <space>s between the 1st two fields in your sample input to 1 <space> between those fields in the output would make it hard for you to determine the data in those fields. And, please clearly explain why additional <space>s need to be added between some output fields when they were not present in your input file? (I.e., describe the logic that determines when additional <space>s need to be added!)
The number of lines in an awk script has absolutely no effect on whether or not you can pipe data into that script or pipe data out of that script. The only reason to insist on a one line awk script is to make it harder to read (especially for someone who is learning how to use awk or who is using a screen that doesn't wrap long lines of code). Our purpose on this site is to help people learn how to use the tools that are available to them on the systems they're using. Writing obfuscated or otherwise unreadable code does not help us achieve that goal.
Please either remove this requirement or make a MUCH more convincing argument as to why you need a one line solution!
P.S. And please answer the question I asked in post #3:
If the answer to this question is "Yes" there might not be any way to achieve your goal.
Notwithstanding the missing answer to Don Cragun's open questions, and for exactly the sample given in post#1, this might come close to a solution:
awk '{LEAD = $1 FS $2 FS $3; TRAIL = $(NF-2) FS $(NF-1) FS $(NF); gsub (/^[^(]*\(|\)[^)]*$/, ""); print LEAD, $0, TRAIL}' file
1234 84 test1 bankUser 10 40 xys
1356 2 branches dirlisting:1 20 80 abc