[SOLVED] How to replace a Request value after a specific pattern in text file?

Issue

Solaris , VERSION: 11.10.0,REV=2005.01.21.15.53
I have a file test.txt which contains values like below:

 <Info>
     <AccountNumber>23456789</AccountNumber>
     <BranchNumber>004</BranchNumber>
     <TransitNumber>01646</TransitNumber>
     <NameAndCity>XYZ Bank</NameAndCity>
     <OwnerFullName>ABC XYZ</OwnerFullName>
  </Info>

All the info is in a line and we have multiple lines like above also other tags are available.

It contains other tag values as well. Also if the tag values contains "333" combination, I don't want to replace them.

I want to use sed command to replace value of the tag with 33333 and after replacement, I want to save the updated info to same file.
Output should be :

 <Info>
     <AccountNumber>33333333</AccountNumber>
     <BranchNumber>33333</BranchNumber>
     <TransitNumber>3333333</TransitNumber>
     <NameAndCity>333 33333</NameAndCity>
     <OwnerFullName>3333 33333</OwnerFullName>
  </Info>

I am new to shell script and not exactly able to write the pattern to match it.

Here is what i have implemented so far to first two tag values but it is not working:

sed 's/(<AccountNumber>)\+[0-2,4-9]*$/\1 33333333/' test.txt
sed 's/(<BranchNumber>)\+[0-2,4-9]*$/\1 33333/' test.txt

Any help will be appreciated .

Solution

Try this:

sed -e '/333/!{' -e 's#<AccountNumber>[0-9]*</AccountNumber>#<AccountNumber>333333333</AccountNumber>#;s#<BranchNumber>[0-9]*</BranchNumber>#<BranchNumber>33333</BranchNumber>#;}'

Eg:

$ sed -e '/333/!{' -e 's#<AccountNumber>[0-9]*</AccountNumber>#<AccountNumber>333333333</AccountNumber>#;s#<BranchNumber>[0-9]*</BranchNumber>#<BranchNumber>33333</BranchNumber>#;}'  test.txt
 <Info>
     <AccountNumber>333333333</AccountNumber>
     <BranchNumber>33333</BranchNumber>
     <TransitNumber>01646</TransitNumber>
     <NameAndCity>XYZ Bank</NameAndCity>
     <OwnerFullName>ABC XYZ</OwnerFullName>
  </Info>

A very straightforward way, if you test ok and want to change the file inplace, add -i switch.

I don't have Solaris to test, so can't be sure.

Try this simple perl one see if it's working:

perl -pe 's#<AccountNumber>[0-9]*</AccountNumber>#<AccountNumber>333333333</AccountNumber>#' test.txt

If it's working, we can add others.

So for your logic first wrote in the question, it should be like this:

perl -pe 'unless (/333/) {s#<AccountNumber>[0-9]*</AccountNumber>#<AccountNumber>333333333</AccountNumber>#;s#<BranchNumber>[0-9]*</BranchNumber>#<BranchNumber>33333</BranchNumber>#;}' test.txt

You can add other substitudes yourself. the # is to replace the usual / of s, an easier way to avoid escape the / in close tags (I.E. s#from#to#;).
It's rather straightforward so I think you'll have no difficulty :)
Add -i switch to change inplace, like this: perl -i -pe '....

Answered By - Tiw