How to extract adjacent strings from txt file with multiple conditions using awk?

I have this txt file

[23/10/10 14:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Request session="lkjh" id=12321>
<type>Old</type>
</Request>
[23/10/10 15:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Request session="lkjhab" id=432>
<type>New</type>
</Request>
[23/10/10 16:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Response session="lkjh" id=12321>
<type>Old</type>
</Response>

I need to use awk to get all requests and responses that have id=12321 AND type "Old".I've never used awk before and i can't find a way to get adjacent strings to the string with id.

The only way i managed to get multiple lines was with grep but only with one pattern.

$ grep id=12321 file.txt -B2 -A2
[23/10/10 14:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Request session="lkjh" id=12321>
<type>Old</type>
</Request>
--
[23/10/10 16:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Response session="lkjh" id=12321>
<type>Old</type>
</Response>

But with grep i can't get requests and responses that have BOTH id=12321 AND type "Old".

Maybe i'm taking wrong approach? Any help would be much appreciated.

Using gnu-awk you can set RS variable to </Request> or </Response> as record separator and then check for 2 search terms in $0:

awk -v RS='</Re(quest|sponse)>' '/id=12321/ && /<type>Old/ {print $0 RT}' file

[23/10/10 14:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Request session="lkjh" id=12321>
<type>Old</type>
</Request>

[23/10/10 16:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Response session="lkjh" id=12321>
<type>Old</type>
</Response>

Like this, with a proper xml parser: [tag:xidel]:

$ xidel -s --input-format=text file.txt -e '
    for $x in tokenize($raw,"\[.+\]  DEBUG")[.]
    return parse-xml($x)[./*[@id=12321 and type="Old"]]
' --output-node-format=xml --output-node-indent 

Credits to Reino

Output

<Request session="lkjh" id="12321">
  <type>Old</type>
</Request>
<Response session="lkjh" id="12321">
  <type>Old</type>
</Response>

A common solution is to set the record separator RS to something which uniquely identifies a new record, so that the current record in each iteration consists of all the lines you want to examine (one entry or related sequence; your target "thing"). Your test data didn't contain any literal square brackets so this is a simple demonstration which works for your sample data:

$ awk 'BEGIN { RS="[" } NR>1 && /id=12321/ && /<type>Old<\/type>/ { print "[" $0 }' <<\:
> [23/10/10 14:37:44:527 EST]  DEBUG
> <?xml version="1.1" encoding="UTF-8" ?>
> <Request session="lkjh" id=12321>
> <type>Old</type>
> </Request>
> [23/10/10 15:37:44:527 EST]  DEBUG
> <?xml version="1.1" encoding="UTF-8" ?>
> <Request session="lkjhab" id=432>
> <type>New</type>
> </Request>
> [23/10/10 16:37:44:527 EST]  DEBUG
> <?xml version="1.1" encoding="UTF-8" ?>
> <Response session="lkjh" id=12321>
> <type>Old</type>
> </Response>
> :
[23/10/10 14:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Request session="lkjh" id=12321>
<type>Old</type>
</Request>

[23/10/10 16:37:44:527 EST]  DEBUG
<?xml version="1.1" encoding="UTF-8" ?>
<Response session="lkjh" id=12321>
<type>Old</type>
</Response>

If you need to accommodate literal square brackets in the data as well, you might perhaps sacrifice the separator line (the one with the square brackets and the DEBUG) and use a regex which uses the entire line as the separator; but that then means that the contents of that line will be discarded as a separator, and not included in the output. (You'll notice that my code above adds back the [ which was "eaten" as a separator.)

With your shown samples in any version of awk please try following code. Written and tested with shown samples Only.

awk '
/^\[[0-9]{2}\/[0-9]{2}\/[0-9]{2}/{
  if(flag2){
    print value
  }
  flag1=flag2=value=""
}
{
  value=(value?value ORS:"") $0
}
/ id=12321>/{
  flag1=1
  next
}
/<type>Old<\/type>/ && flag1{
  flag2=1
}
END{
  if(flag2){
    print value
  }
}
'   Input_file