Print a pattern between the xml tags based on a search pattern

Shell Programming and Scripting
1

Hi all,

I am trying to extract the values ( text between the xml tags) based on the Order Number.

here is the sample input

<?xml version="1.0" encoding="UTF-8"?>
<NJCustomer>
    <Header>
        <MessageIdentifier>Y504173382</MessageIdentifier>
        <ProcessIdentifier>253128</ProcessIdentifier>
        <MessageProducer>NJCustomer ERP ISF</MessageProducer>
        <MessageConsumer>NJ</MessageConsumer>
        <MessageFunction>ErpOrderNotification</MessageFunction>
        <MessageDateTime gmtOffset="-4">2011-03-18T06:01:43.209-04:00</MessageDateTime>
    </Header>
    <ErpOrderNotification type="Booked">
        <OrderNumber>939511</OrderNumber>
        <ErpOrderNumber>504173382</ErpOrderNumber>
        <ErpOrderStatus>Booked</ErpOrderStatus>
        <StatusChangeDateTime gmtOffset="-04:00">20110318 0601</StatusChangeDateTime>
        lOfServiceCode>LOCP</LevelOfServiceCode>
        <CarrierCode>LOCP</CarrierCode>
        <Location type="destination">
            <LocationCode>c/o Arvato Distribution GmbH</LocationCode>
        </Location>
    </ErpOrderNotification>
</NJCustomer>

I need to feed the order number (Here in this example 939511) and it should display the text between <NJCustomer> and </NJCustomer>.

The input file is very large and i feel a solution in awk could be better.

I searched the forum and i got this code and seems it needs small modification and i am not sure what this code does.

awk -v order=$ORD '/<NJCustomer>/{if(l)print s;l=0;s=$0;next}/order/{l=1}{s=s RS $0}END{if(l)print s}' <filename>
2

For XML processing I'd prefer Perl:

perl -ln0e 'while (/<NJCustomer>.*?<\/NJCustomer>/gs){$x=$&;print $x if $x=~/<OrderNumber>939511<\/OrderNumber>/}' file
3

Bartus...yes its works and difficult to understand the perl line. But how to change the order number as variable(mean for different order number)

4

If order number is stored in shell variable ORD, then this will work:

perl -ln0e "while (/<NJCustomer>.*?<\/NJCustomer>/gs){\$x=\$&;print \$x if \$x=~/<OrderNumber>$ORD<\/OrderNumber>/}" file
1 Like
5 
awk -F'[<|>]' '$2 ~ "NJCustomer"{f=f?0:1}f && $2 =="OrderNumber"{print $3}' file
6

Danmero,

Your solution reveals the order number only...but my query is if the order number is there in the file corresponding <NJCustomer> & </NJCustomer> needs to be printed

7

Ups...:rolleyes:

awk -F'[<|>]' '$2 ~ "NJCustomer"{f=f?0:1}f && $2 =="OrderNumber"{print $3}' file
8

f=f?0:1 could be shorten f=!f

9

Danmero & ctsgnb,

Its the same thing...actually the perl code given by bartus works fine...but an if just looking for an awk solution(since the non-availability of the perl in the current environment).

10

Oky, Could this help you?

awk '/<NJCustomer>/ , /<\/NJCustomer>/{if($0~/<OrderNumber>'$ord'<\/OrderNumber>/){print str;print $0;flg=1;next}else{str=str"\n"$0}if(flg==1){if(/<NJCustomer>/){str="";flg=0}else{print}}}' input.xml
1 Like
11

That's correct ... only if you want to print only the first occurrence.
In this case we should exit after print

awk -F'[<|>]' '$2 ~ "NJCustomer"{f=1}f && $2 =="OrderNumber"{print $3;exit}' file
12

pravin,

Ur one liner works like a charm..but i could understand only half of it...if u explain it will be gr8...

13

Oky, Could this help you?

awk '/<NJCustomer>/ , /<\/NJCustomer>/  # We are taking text from input xml which is between these two tags.
{if($0~/<OrderNumber>'$ord'<\/OrderNumber>/) # If any of the text(which we filtered in previous statement) match the string "<OrderNumber>'$ord'<\/OrderNumber>" ($ord is your order number) then
{print str;print $0;flg=1;next} # print str(i.e text prior to string "<OrderNumber>'$ord'<\/OrderNumber>" and current line i.e. "<OrderNumber>'$ord'<\/OrderNumber>"
else{str=str"\n"$0} # here we are taking prior text.
if(flg==1)
{if(/<NJCustomer>/) {str="";flg=0} #if my current  record is /<NJCustomer>/ means start of new section then reset the "str" make flag zero.
else{print}}}' # if my flag is one then print the current record .i.e. text after string "<OrderNumber>'$ord'<\/OrderNumber>"
14

Pravin.....Thanks