Shell Programming and Scripting

Hello RudiC,

Thank you for your reply! It doesn't work for me. I assume that the grep command you gave me is missing the XML file from where the information should be extracted.

Milano

You assume incorrectly. The code RudiC provided does exactly what you asked for given the filenames you provided. But, of course we're making assumptions about the utilities you have installed on your system, the shell you're using, and the operating system you're using.

What operating system are you using?
What version of UNIX/Linux utilities are you using?
What shell are you using?
What output did RudiC's code produce on your system?
Are you sure that the filenames you provided contain data in the same format as your sample data? (For instance, does C:/temp/input.txt contain <carriage-return><newline> line terminators instead of the <newline> line terminators expected by UNIX and Linux system utilities?)

Yes, this was my mistake. Is working good for my example. I will test it on larger files too. Now I am trying to delete the extracted rows from the original XML. Hope I will manage it.
Thnak you!

Milano

---------- Post updated at 05:19 AM ---------- Previous update was at 02:28 AM ----------

I tried this to remove the line that were extracted from the xml file, but I got an error of

Code:

enexpected end of file

when I am trying to run the script.

Code:

sed -i 'iE "$(tr -d "'" </home/qqomtws/kwom/Test_HY/test_hy.txt | tr '\n' '|')" /home/qqomtws/kwom/Test_HY/test_hy.xml' ./infile

I put the command just after the grep command that extracts the lines I am interested in. Any advice?

Thank you!
Milano

---------- Post updated 06-19-15 at 03:07 AM ---------- Previous update was 06-18-15 at 05:19 AM ----------

Yes, this was my mistake. Is working good for my example. I will test it on larger files too. Now I am trying to delete the extracted rows from the original XML. Hope I will manage it.
Thnak you!

I tried this to remove the lines that were extracted from the xml file, but I got an error of

Code:

unexpected end of file

when I am trying to run the script.

Code:

sed -i 'iE "$(tr -d "'" <C:/temp/input.txt | tr '\n' '|') " C:/temp/output.txt'./infile

I put the command just after the grep command that extracts the lines I am interested in. Any advice?

Milano

---------- Post updated at 04:22 AM ---------- Previous update was at 03:07 AM ----------

Yes, this was my mistake. Is working good for my example. I will test it on larger files too. Now I am trying to delete the extracted rows from the original file. Hope I will manage it.
Thnak you!

I tried this to remove the line that were extracted from the xml file, but I got an error of

Code:

enexpected end of file

when I am trying to run the script.

Code:

sed -i 'iE "$(tr -d "'" <C:/temp/input.txt | tr '\n' '|') " C:/temp/output.txt'./infile

I put the command just after the grep command that extracts the lines I am interested in. Any advice?

Many thanks,
Milano

You can't do that, replace one comand ( grep ) by another ( sed ) with identical parameter set, and hope that it works.
To remove those selected lines from the original file, redirect the grep result to a temp file and try

Code:

grep -vfTMP C:/temp/output.txt

The obvious simple thing to do (to extract all of the lines that the 1st grep did NOT extract) would be to just rerun that script adding a -v option:

Code:

grep -viE "$(tr -d "'" <C:/temp/input.txt | tr '\n' '|')header|footer" C:/temp/output.txt

Or, you could use RudiC's suggestion (but I would add a -F in case some of the strings extracted by the first grep contain characters that are special in a BRE):

Code:

grep -vFfTMP C:/temp/output.txt

If you are going to be running this script regularly, I would seriously consider rewriting it to use awk instead of grep. If you use awk you could produce both output files in one pass without needing to run grep twice and without needing to read your input XML file twice:

Code:

awk -v ERE="$(tr -d "'" <C:/temp/input.txt | tr '\n' '|')header|footer" '
BEGIN {ERE = tolower(ERE)}
      {print > ((tolower($0) ~ ERE) ? "matched.xml" : "unmatched.xml")}
' C:/temp/output.txt

Change matched.xml and unmatched.xml to the pathnames of the files you want to contain the matched lines and the unmatched lines, respectively. I assume that you already know that neither of those output files can be the input file for this awk script!

Thank you a lot! It works very well! A great advice also with

Code:

awk

Many thanks,
Milano