sed get specific pattern and change it

Question

I have file with following content

<general id="7" src="xyz" srcType="abc" link="positive" />
<general id="3" src="xyz" srcType="pqr" link="positive" />
..

I am trying to replace every src name with its capitalize word. For example in above I want to replace all xyz with Xyz (first letter is capital).

Expected output is

<general id="7" src="Xyz" srcType="abc" link="positive" />
<general id="3" src="Xyz" srcType="pqr" link="positive" />
..

I tried regex pattern src="(.*?)" which can extract 'xyz' from the line but how can I use it with sed to replace in my file?

I tried following but it is not working

sed -E 's/srcName="(.*?)"/Xyz/1' filename

Bonus question, how can I generalize it? Where anything between scr=" and " can be converted into word capitalizing first character.

Answer 1

To edit XML, use an XML-aware tool. For example, there's xsh, a tool I adopted to maintain.

open file.xml ;
for //@src set . xsh:subst(., "(.)", "\U$1") ;
save :b ;

Answer 2

A XML-aware tool, is definitely the way! For a quick hack:

in GNU sed:

sed -r 's/src="([a-z])/src="\U\1\E/g'  file.xml

for more complex changes I suggest Perl

perl -pe 's/src="\K(\w+)/ucfirst($1)/e'   file.xml

where

• perl -pe command is a "readline, command, print" loop
• s/regexp/expression/e substitutes regexp by the result of eval(expression)
• src="\K is a left-context pattern

Answer 3

Tried with awk command and it worked fine

awk -F 'src="' '{$2=(toupper(substr($2,1,1))substr($2,2));$2="src="$2;print $0}' filename|sed 's/src=/&"/g'

output

<general id="7"  src="Xyz" srcType="abc" link="positive" />
<general id="3"  src="Xyz" srcType="pqr" link="positive" />