Process an XML File with AWK

[IMAUser]

Hi ,
Have an XML file a section of which is as below
<component>
<refentity>
<name>ABITIBI-CONSOLIDATED INC.</name>
<red>003CB6</red>
</refentity>
<bond>
<name>ABY 8.55 01Aug10</name>
<issuer>
<prospname>Abitibi</prospname>
<pairred>003CB6AA3</pairred>
<ispreferred>true</ispreferred>
</issuer>
<type>Bond</type>
<isconvert>false</isconvert>
</bond>
<red>003CB6AA3</red>
</component>
<component>
<refentity>
<name>Argentine Republic</name>
<red>PP7D7E</red>
</refentity>
<bond>
<name>ARGENT 8.28 31Dec33 Sink</name>
<type>Bond</type>
<isconvert>false</isconvert>
</bond>
</component>

The above are two <component> records. In the first one within the tag <bond> there is a tag <issuer> , in the second there is no <issuer>.

I need to write a script to seperately write all the components and bonds which have a issuer into one file and components and bonds which do not have an issuer into another file.

Any help is very much appreciated.

Thanks

 

[feherke]

Hi

awk '
/<component>/,/<\/component>/ { s=s ORS $0 }
/<\/component>/ { print s > (s~/<issuer>/?"with":"without") "_issuer"; s="" }
' inputfile.xml

Feherke.

http://rootshell.be/~feherke/

 

[IMAUser]

Thanx for that but it complains saying
awk: syntax error near line 3
awk: illegal statement near line 3

If you can tell me what each bit is doing there, I may be able to debug it.

Thanks again.

 

[feherke]

Hi

I have [tt]gawk[/tt] 3.1.1 and works fine. Works with --traditional option too, so theoretically should be compatible with other [tt]awk[/tt] versions.

Try to put all the filename expression into parenthesis :

awk '
/<component>/,/<\/component>/ { s=s ORS $0 }
/<\/component>/ { print s > [red]([/red](s~/<issuer>/?"with":"without") "_issuer"[red])[/red]; s="" }
' inputfile.xml

All it does, is to put all lines found between component marks into a variable, then when the end of section is reached, prints that value in a file.

/<component>/,/<\/component>/ {   [gray]# for lines between start and end marks[/gray]
  s=s ORS $0                      [gray]# add the current line to the variable[/gray]
}
/<\/component>/ {                 [gray]# the end mark is reached[/gray]
  print s >                       [gray]# print the variable's value to the file...[/gray]
    ((s~/<issuer>/?"with":"without") "_issuer")
[gray]# *_issuer, where the prefix of "with" or "without" is determined by the presence of issuer mark in value of s[/gray]
  s=""                            [gray]# reset the accumulated value[/gray]
}

Feherke.

http://rootshell.be/~feherke/