[Sudarshan_Kanna]

<ARS>

<tag1>one</tag1>

<tag2>two</tag2>

<tag3>three<AltError Code=123 Description=456789/></tag3>

<tag4>four</tag4>

<ARS>

<ARS>

<tag1>ABCD</tag1>

<tag2>ABCD</tag2>

<tag3>ddsdsds<AltError Code=123 Description=456789/></tag3>

<tag4>EFGH<AltError Code=abc Description=defg/></tag4>

<ARS>



Expected Output:



tag1|tag2|tag3|tag4|code|Description

one|two|three|four|123|456789

ABCD|ABCD|ddsdsds|123|456789

ABCD|ABCD|ddsdsds|abc|defg

The script should read this file and create a pipe delimited file. Each ARS tag record in the XML file should create one line in the target. Get all the values enclosed within the tags, which are the column values.



Also if suppose an <AtlError> encounters within an ARS tag, then get the Code and Description values of the AltError tag and append it to the last of that particular ARS record. If suppose 2 AltError encounters within an ARS tag, the target should be having 2 records

[Simon]

You should get acquainted with sed.

[Gee4orce]

Sure can.

My normal rate is $70/hour

[rehoot]

I have a script that is close to what you want, but it lease the newlines in the output. I think the special characters will cause a problem when I post the script: the slanted smiley faces are equals signs.

This is all one line:

cat ~/input.txt|sed -e s'/\(<tag[0-9]>\)\([^<]*\)<\/tag[0-9]>/\2|/g' -e s'/ \{1,30\}//g' -e s'/<tag[0-9]>\([^<]*\)<AltError[ ]*Code=\([^ ]*\)[ ]*Description=\([^ ]*\)/\1|Code1|Description2/g' -e s'/<ARS>//g'

Some notes:
1) the 'cat' program just prints the file
2) a block in a sed command that is found using a string that is surrounded with \( and \) can later be displayed using \1 (or \2 if it is the second occurrence).
3) the code that says \([^ ]*\) means to find a bunch of characters until you reac a space (the carret means "NOT" and it is follwed by a space).
4) I added the strings: "Code:" and "Description:" to the output
5) tag[0-9] finds the string "tag" followed by a number from 0 to 9. Add a star after this if there are more digits.