Reputation: 17
I have an xml files and I want to fetch the values of some tags. XML is something like this given below:
<?xml version="1.0" standalone = "no"?>
<!DOCTYPE handover_list PUBLIC"EN""h">
<X1>
<X2>
<X3>USA</X3>
<date_time>20170813T18:18-04:00</date_time>
<id action="I">XXXXXXXXXXXXXX</id>
<id action="I">YYYYYYYYYYYYYY</id>
<id action="I">ZZZZZZZZZZZZZZ</id>
</X2>
<X2>
<X3>UAE</X3>
<date_time>20160814T15:15-03:04</date_time>
<id action="I">AAAAAAAAAAAAAA</id>
<id action="I">BBBBBBBBBBBBBB</id>
<id action="I">CCCCCCCCCCCCCC</id>
</X2>
</X1>
What I'm using is:
sed -n 's:.*<X3>\(.*\)</X3>.*:\1:p' formated.xml
sed -n 's:.*<id action="I">\(.*\)</id>.*:\1:p' formated.xml
and its given output like this:
USA
UAE
XXXXXXXXXXXXXX
YYYYYYYYYYYYYY
ZZZZZZZZZZZZZZ
AAAAAAAAAAAAAA
BBBBBBBBBBBBBB
CCCCCCCCCCCCCC
What I want is to merge both the sed commands used above so that I can get the output like this:
USA
XXXXXXXXXXXXXX
YYYYYYYYYYYYYY
ZZZZZZZZZZZZZZ
UAE
AAAAAAAAAAAAAA
BBBBBBBBBBBBBB
CCCCCCCCCCCCCC
Upvotes: 0
Views: 253
Reputation: 433
GNU sed:
sed '/<X3>/{s/<[^>]*>//g;h};/ action=/{s/<[^>]*>//g;H};/<\/X2>/{g;b};d' formated.xml
Upvotes: 0
Reputation: 92854
The right way is using xml parsers like xmlstarlet:
In such case, <DOCTYPE ..>
tag is redundant.
xmlstarlet sel -t -v '//X2/*[not(self::date_time)]' -n formated.xml
The output:
USA
XXXXXXXXXXXXXX
YYYYYYYYYYYYYY
ZZZZZZZZZZZZZZ
UAE
AAAAAAAAAAAAAA
BBBBBBBBBBBBBB
CCCCCCCCCCCCCC
Upvotes: 1
Reputation: 88583
Concatenate both sed commands with one ;
:
sed -n 's:.*<X3>\(.*\)</X3>.*:\1:p' formated.xml
sed -n 's:.*<id action="I">\(.*\)</id>.*:\1:p' formated.xml
To one sed command:
sed -n 's:.*<X3>\(.*\)</X3>.*:\1:p; s:.*<id action="I">\(.*\)</id>.*:\1:p' formated.xml
Output:
USA XXXXXXXXXXXXXX YYYYYYYYYYYYYY ZZZZZZZZZZZZZZ UAE AAAAAAAAAAAAAA BBBBBBBBBBBBBB CCCCCCCCCCCCCC
Upvotes: 1