Help needed in writing awk script for xml source

naren_0101bits · 09-23-2007, 12:46 AM

Hi, i am not able to get an approach for converting xml file to flat file using awk programming. Can anyone help me out.

The input xml is like this:

<outer>
<field1>one</field1>
<field2>two</field2>
<field3>three<Error Code=777 Description=12345/></field3>
<field4>four</field4>
</outer>

Expected output is like this:

field1 field2 field3 field4 Code Description
one two three four 777 12345
Thanks in advance,
Naren

radoulov · 09-23-2007, 04:30 AM

Code:

 awk 'FNR==1{printf "%-6s|%-6s|%-6s|%-6s|%-6s|%s\n","field1","field2", 
"field3","field4","code","description"}
/field1/{printf "%-6s|",$3}
/field2/{printf "%-6s|",$3}
/field3/{printf "%-6s|",$3;code=$6;desc=$8}
/field4/{printf "%-6s|%-6s|%s\n",$3,code,desc}' FS="</|[/ <>=]" filename

Example:

Code:

zsh 4.3.4% cat file
<outer>
<field1>one</field1>
<field2>two</field2>
<field3>three<Error Code=777 Description=12345/></field3>
<field4>four</field4>
</outer>
<outer1>
<field1>one1</field1>
<field2>two1</field2>
<field3>three1<Error Code=7770 Description=1234/></field3>
<field4>four1</field4>
</outer1>
<outer2>
<field1>one2</field1>
<field2>two2</field2>
<field3>three2<Error Code=7 Description=45/></field3>
<field4>four2</field4>
</outer2>
zsh 4.3.4% awk 'FNR==1{printf "%-6s|%-6s|%-6s|%-6s|%-6s|%s\n","field1","field2", 
"field3","field4","code","description"}
/field1/{printf "%-6s|",$3}
/field2/{printf "%-6s|",$3}
/field3/{printf "%-6s|",$3;code=$6;desc=$8}
/field4/{printf "%-6s|%-6s|%s\n",$3,code,desc}' FS="</|[/ <>=]" file
field1|field2|field3|field4|code  |description
one   |two   |three |four  |777   |12345
one1  |two1  |three1|four1 |7770  |1234
one2  |two2  |three2|four2 |7     |45

Note that this is only an example code and
it's quite possible that it doesn't work with your real data (you may have to adjust the FS, use sub/gsub etc.).
If you post a real sample from you xml,
it will be better.

P.S. And of course, if you install the XML gawk (xgawk) you won't need this "ugly" code

naren_0101bits · 09-24-2007, 12:12 AM

Is it possible to do the same with out hardcoding the field values field1, field2, field3, code, description except the outer tag.
what if there is one more level of inner tags like wise
<outer>
<field1>one</field1>
<field2>two</field2>
<field3>three<Error Code=777 Description=12345/></field3>
<field4>four</field4>
<inner>
<f1>1</f1>
<f2>2</f2>
</inner>
</outer>

ghostdog74 · 09-24-2007, 12:21 AM

don't waste time and keep changing requirements, show the whole input sample xml file you have if possible. describe clearly what you want as output.

naren_0101bits · 09-24-2007, 12:36 AM

naren_0101bits · 09-26-2007, 07:48 AM

Refresh
Can anyone suggest me some approach for my problem

matthewg42 · 09-26-2007, 08:00 AM

You could use the program xmlstarlet.

You could also use Perl and the XML::Parser module.

angrybanana · 10-02-2007, 08:47 PM

Bit late, but here's a python solution:

Code:

$ cat a
<AMP>
<EnrID>ENR1</EnrID>
<SubID>SUB135</SubID>
<DependentID>0011</DependentID>
<StfID>Sysuser<Error Code="233" Description="Staff ID does not exist"/></StfID>
<PRoc>
<ProcCType>ICU</ProcCType>
<ProcCode>31</ProcCode>
<Rank>SEC</Rank>
<Active></Active>
<InActive>12319999</InActive>
</PRoc>
</AMP>

$ cat a|python parser.py
EnrID|SubID|DependentID|StfID|Code|Description|ProcCType|ProcCode|Rank|Active|InActive
ENR1|SUB135|0011|Sysuser|233|Staff ID does not exist|ICU|31|SEC|-|12319999

here's the script:

Code:

import sys
from xml.etree.ElementTree import ElementTree

d = {}
tags = []
tree = ElementTree(file=sys.stdin)

for element in tree.getiterator():
        if not element.attrib and element.text=='\n':
                continue
        if element.tag=="Error":
                d.update(element.attrib)
                tags += element.attrib.keys()
        else:   
                element.text = element.text or '-'
                d[element.tag] = element.text.strip()
                tags.append(element.tag)

print '|'.join(tags)
print '|'.join(d[tag] for tag in tags)

Hope this helps.