XML Fields comparison using awk script

Shell Programming and Scripting
1

Hello All,

I have many zipped XMLs (example file name in tgz formate - file_rec.trx.2016-01-23.000123.exc.85sesdzd45wsds5299c8f2994f7.tgz ) looks following and I need to verify two numbers, they are RecordNumber and EnrolData (only sequence number, NOT hole) .
for all the records, both should be equal, but as an error, for some records, record number is NOT same as EnrolData's sequence number. I need to find out what all those records and in which files. could some one please help me? I have tried this using following awk script but no luck.

XML Format:

<XXXXXXXXXXXXX>
    <RecordNumber>12345</RecordNumber>
    <XXXXXX>XXXXXX</XXXXXX>
    <XXXXXX>XXXXXX</XXXXXX>
    <XXXXXX>XXXXXX</XXXXXX>
    <XXXXXXXXXXXXX><![CDATA[XXXXXXXXXXXXXX:XXXXXXXXXXXXX XXXX XXXXXX]]></XXXXXXXXXXXXX>
    <EnrolData><![CDATA[E0000003350000000012345Part1              XXXXXX
	XXXXXXXXXXXXXXXX                                            XXXXXXXXXXXXXXX:XXXXXXXXXXXXXXXXXXXXXXXXXXX.XXXXXXXXXXXXXXXXXXXXXXXXXXX.XXX   
	XXXXXXXXXXXXXXX  
	XXXX                                                                                                                                                      
	
XXXXXXXXXXXXXXXXX                    XXXX                                XXXXXXXXXXXXX.XXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXX.XXXXXX                                                        XXXXXXXXXXXXXX                                      
        XXXX                          XXXX                          XXXXXXXXXXXXX                XXXX                          XXXXXXXXXXXXX       
		XXXX                          XXXXXXXXXXXXX                X                            
		XXXXXXXXXXXXX                                                                                             		
		XXXXXXXXXXXXXXXXXXX.XXXXXXXXXXXXXXXXXXXX.XXX                           XXX
]]></EnrolData>
</XXXXXXXXXXXXX>

Script that I am trying:

#!/bin/sh
for file in $(ls file_rec.trx.{4}(\d)-{2}(\d)-{2}(\d).{1,}(\d).exc.*.tgz)
do
awk'
 /<RecordNumber>/ {
        getline
        while ( $0 !~ /<\/RecordNumber>/ ) {
               rNumber = $1
                getline
        }
        nextline
}

/<EnrolData><\!\[CDATA\[/ {
        getline
        while ($0 !~ "\]\]><\/EnrolData>" ) {
               eData=substr($1,19,5) #Here I actually need to get the sub string from "E0000003350000000012345Part1              XXXXXX                                        " 
#but the problem is record number may not fixed digits and the number between Part1 and E may not be fixed digits. 
#one thing for sure is sequence number present always before Part1
                getline
        }
        nextline
}
{
if (rNumber==eData){
#here I need to print the formate - <filename> : <RecordNumber> - <EnrolData sequence number>
print "$file - $(rNumber) - $(eData)"
}' $file
2

So - what be the EnrolData? If it's NOT 3350000000012345 - what is it?

3

Hello Rudi, Thank you for checking this for me. here is the answer:

E00000033500000000 - Some different string which you may ignore. unfortunately this string length may change.
12345 - this is the actual sequence number that we need to compare with RecordNumber. In other words, this is the sequence number which must be equal to the Record number
Part1 - This is also another string and is FIXED for each and all files.

---------- Post updated at 01:39 PM ---------- Previous update was at 01:28 PM ----------

To be more clear, in below XML, the two separated numbers (12345 in two XML tags - RecordNumber and EnrolData) must be equal, but for some reason, in some records, they are not coming as same. Also the string Part1 is same for all records and for all files.
So I need to find out in what file how they are not coming as different.

<XXXXXXXXXXXXX>
    <RecordNumber>

12345

</RecordNumber>
    <XXXXXX>XXXXXX</XXXXXX>
    <XXXXXX>XXXXXX</XXXXXX>
    <XXXXXX>XXXXXX</XXXXXX>
    <XXXXXXXXXXXXX><![CDATA[XXXXXXXXXXXXXX:XXXXXXXXXXXXX XXXX XXXXXX]]></XXXXXXXXXXXXX>
    <EnrolData><![CDATA[E00000033500000000

12345

Part1              XXXXXX
	XXXXXXXXXXXXXXXX                                            XXXXXXXXXXXXXXX:XXXXXXXXXXXXXXXXXXXXXXXXXXX.XXXXXXXXXXXXXXXXXXXXXXXXXXX.XXX   
	XXXXXXXXXXXXXXX  
	XXXX                                                                                                                                                      
	
XXXXXXXXXXXXXXXXX                    XXXX                                XXXXXXXXXXXXX.XXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXX.XXXXXX                                                        XXXXXXXXXXXXXX                                      
        XXXX                          XXXX                          XXXXXXXXXXXXX                XXXX                          XXXXXXXXXXXXX       
		XXXX                          XXXXXXXXXXXXX                X                            
		XXXXXXXXXXXXX                                                                                             		
		XXXXXXXXXXXXXXXXXXX.XXXXXXXXXXXXXXXXXXXX.XXX                           XXX
]]></EnrolData>
</XXXXXXXXXXXXX>

Hope this clarifies.

4

So - would 102345 and 10002345 or even 10002 be valid numbers? If so, how to discriminate from above ( 3350...012345 )?

---------- Post updated at 10:44 ---------- Previous update was at 10:43 ----------

We need something to tell where to stop ignoring digits and start counting them...

5

Hello RudiC, unfortunately that is what the problem for me, there is no valid scenario for this, Record number length is not fixed, but the number right before Part1 is the number that we need to compare with the Record Number. I am not sure how to extract this.

6

Well, assuming the record number length is as good an approach as any other, try

awk '
FNR == 1                                {RN = ED = 0}
match ($0, /<RecordNumber>[^<]*/)       {RN=substr($0, RSTART+14, RLENGTH-14); print FILENAME; print RN}
match ($0, /<EnrolData>.*Part1/)        {ED=substr($0, RSTART+RLENGTH-5-length(RN), length(RN)); print ED}
' file
file
12345
12345