value on some lines but no others, extracting field problem 1

[LGJ]

Please help:

I have a data file which has many values per line eg:

lemur/front-stats-2003102459:<apiCall start=&quot;1066950214507&quot; end=&quot;1066950216147&quot; remoteHost=&quot;xxx&quot; remoteAddr=&quot;xxx&quot; contractID=&quot;oranlc1&quot; serviceName=&quot;BillingServices.Content.ChargeReservation&quot;><apiResponse success=&quot;yes&quot;></apiResponse>
</apiCall>
lemur/front-stats-2003102459:<apiCall start=&quot;1066950219273&quot; end=&quot;1066950219739&quot; remoteHost=&quot;xxx&quot; remoteAddr=&quot;xxx&quot; contractID=&quot;oranlc1&quot; serviceName=&quot;BillingServices.Content.ChargeConfirmation&quot;><apiResponse success=&quot;yes&quot;></apiResponse
></apiCall>
lemur/front-stats-2003102459:<apiCall start=&quot;1066950219903&quot; end=&quot;1066950225208&quot; remoteHost=&quot;xxx&quot; remoteAddr=&quot;xxx&quot; contractID=&quot;akumiitti2orangelc1&quot; serviceName=&quot;BillingServices.Content.ChargeConfirmation&quot;><apiResponse success=&quot;no&quot;><reason><code>
107</code></reason></apiResponse></apiCall>

the above shows 3 lines (start with lemur, end with </apiCall>)

Originally I only wanted the start_t, end_t, charge_request and the success value so the following statement was sufficient:

awk -F\&quot; '{print $2&quot;,&quot;$4&quot;,&quot;$12&quot;,&quot;$14}' $InputFile > $OutputFile

I now need the code value if the success = no (between <code> and </code>) but this is not between double quotes AND is only on the failed transaction lines eg 3rd line of given example.

Please can someone suggest a solution to this, it has been racking my brains and I produced a crude solution using sed and awk to replace unwanted text!! Surely there is something in awk to solve this?

Thanks all

Lee

 

[Ygor]

Try this...

awk -F\&quot; '{split($15,a,&quot;[<>]&quot;print $2&quot;,&quot;$4&quot;,&quot;$12&quot;,&quot;$14&quot;,&quot;a[6]}' $InputFile > $OutputFile

 

[aigles]

Try this :

awk -F\&quot; '
{
if ($14 == &quot;no&quot;/) {
code = $0 ;
sub(&quot;.*<code>&quot;,&quot;&quot;,code);
sub(&quot;</code>.*&quot;,&quot;&quot;,code);
code = &quot;,&quot; code;
} else {
code = &quot;&quot;;
}
print $2&quot;,&quot;$4&quot;,&quot;$12&quot;,&quot;$14 code ;
} ' $InputFile > $OutputFile ;


Jean Pierre.

 

[LGJ]

Thanks for your time but I cant seem to get any of these working properly:

Ygor I get this kind of output:

1066950219273,1066950219739,BillingServices.Content.ChargeConfirmation,yes,
1066950219903,1066950225208,BillingServices.Content.ChargeConfirmation,no,/apiResponse>

I have messed around with the array value but still cant get just the code value eg 107

Jean Pierre I get syntax errors being shown, I like this solution as I can understand what is happening but I cannot seem to work out why it wont run? ( i have removed / after no as I dont think this is needed also)

Any ideas

thanks

Lee

 

[aigles]

With / removed, the script works fine on my system (AIX).
The result is :

1066950214507,1066950216147,BillingServices.Content.ChargeReservation,yes
1066950219273,1066950219739,BillingServices.Content.ChargeConfirmation,yes
1066950219903,1066950225208,BillingServices.Content.ChargeConfirmation,no,107

What errors did you get ?

Jean Pierre.

 

[PHV]

And something like this ?
awk '{a=&quot;&quot;;b=&quot;&quot;;c=&quot;&quot;;d=&quot;&quot;}
/start=/{a=$0;sub(/.*start=&quot;/,&quot;&quot;,a);sub(/&quot;.*/,&quot;&quot;,a)}
/end=/{b=$0;sub(/.*end=&quot;/,&quot;&quot;,b);sub(/&quot;.*/,&quot;&quot;,b)}
/serviceName=/{c=$0;sub(/.*serviceName=&quot;/,&quot;&quot;,c);sub(/&quot;.*/,&quot;&quot;,c)}
/<code>/{d=$0;sub(/.*<code>/,&quot;,&quot;,d);sub(/<.*/,&quot;&quot;,d)}
{printf &quot;%s,%s,%s%s\n&quot;,a,b,c,d}
' $InputFile > $OutputFile

Hope This Help
PH.

 

[LGJ]

aigles/Jean Pierre:

I get the following:

pin:/export/home/pin/temp/ltemp> awk -F\&quot; '
{
if ($14 == &quot;no&quot {
code = $0 ;
sub(&quot;.*<code>&quot;,&quot;&quot;,code);
sub(&quot;</code>.*&quot;,&quot;&quot;,code);
code = &quot;,&quot; code;
} else {
code = &quot;&quot;;
}
print $2&quot;,&quot;$4&quot;,&quot;$12&quot;,&quot;$14 code ;
} ' testFile > OutputFile;
awk: syntax error near line 5
awk: illegal statement near line 5
awk: syntax error near line 6
awk: illegal statement near line 6

 

[aigles]

I think you are using an old version of awk which does not support the sub function.

Try to use nawk instead of awk.

Jean Pierre.

 

[LGJ]

cool aigles, it now compiles fine (with nawk) but the output I get is:

1066950214507,1066950216147,BillingServices.Content.ChargeReservation,yes
,,,
1066950219273,1066950219739,BillingServices.Content.ChargeConfirmation,yes
,,,
1066950219903,1066950225208,BillingServices.Content.ChargeConfirmation,no
,,,

The code you produced looks like it should work though?

Thanks

Lee

 

[aigles]

Empty input lines gives &quot;,,,&quot; in output.
Are you sure that your input file contains only valid lines (starting with lemur and ending with </apiCode>.

You can modify the script to select only this lines :

awk -F\&quot; '
/^lemur.*<\/apiCall>$/ {
if ($14 == &quot;no&quot {
code = $0 ;
sub(&quot;.*<code>&quot;,&quot;&quot;,code);
sub(&quot;</code>.*&quot;,&quot;&quot;,code);
code = &quot;,&quot; code;
} else {
code = &quot;&quot;;
}
print $2&quot;,&quot;$4&quot;,&quot;$12&quot;,&quot;$14 code ;
} ' txt

I don't understand why the code isn't displayed for the last line.
The script works fine on my system whith your input example.


Jean Pierre.

 

[LGJ]

Thank you Jean Pierre (Aigles) for your help.

I had concatenated one of the lines and $0 was incomplete for the line with <code> on it!

I have learnt some new skills today.

Cheers

Lee

 

[Ygor]

My earlier post should have looked like this (but it didn't because I left TGML on)...

awk -F\&quot; '{split($15,a,&quot;[<>]&quot;print $2&quot;,&quot;$4&quot;,&quot;$12&quot;,&quot;$14&quot;,&quot;a[6]}' $InputFile > $OutputFile