如何首先grep一个单词，然后如何使用TCL regexp根据该单词grep输出？

Question

This is my TCL Script:

set test {

device#more system:/proc/dataplane/fw/application
1 : Amazon Instant Video (num of policy actions: 0)
     port-proto:
     http urls :
              *(www.amazon.com/Instant-Video)*
     dns names :
     https client-hello servNames :
     https server-hello servNames :
     https server-certificate commonNames :
     Application stats :
             Bytes Uploaded : 0
             Bytes Download : 0
             Num Flows : 0
2 : SIP (num of policy actions: 0)
     port-proto:
             Proto 6-6, sport 0-65535, dport 5060-5061
             Proto 17-17, sport 0-65535, dport 5060-5061
     http urls :
     dns names :
     https client-hello servNames :
     https server-hello servNames :
     https server-certificate commonNames :
     Application stats :
             Bytes Uploaded : 0
             Bytes Download : 0
             Num Flows : 0
3 : Photobucket (num of policy actions: 0)
     port-proto:
     http urls :
              *(www.pbsrc.com)*
              *(www.photobucket.com)*
     dns names :
              *.photobucket.co (2)
              *.photobucket.com (2)
     https client-hello servNames :
     https server-hello servNames :
     https server-certificate commonNames :
     Application stats :
             Bytes Uploaded : 34
             Bytes Download : 44
             Num Flows : 78
4 : Filestub (num of policy actions: 0)
     port-proto:
     http urls :
              *(www.filestub.com)*
     dns names :
              *.filestub.com (2)
     https client-hello servNames :
     https server-hello servNames :
     https server-certificate commonNames :
     Application stats :
             Bytes Uploaded : 0
             Bytes Download : 0
             Num Flows : 0
--More--
device#

}



set lines [split $test \n] ; # split using new line char(\n)
set data [join $lines :]
if { [regexp {Photobucket.*(Bytes Uploaded : .* Bytes Download:)} $data x y]} {
        set y [string trimright $y {: }]
        puts "Bytes uploaded : $y"
    }

I am trying to find the Bytes downloaded and uploaded to the application called "Photobucket" in $test variable.

STEPS that script to do:

1. First identify the word "Photobucket" 
2. Then grep for "Bytes Uploaded : <any number> and Bytes Download : <any number>, Num Flows : <any number> for that application "Photobucket".

Output should be:

Application Name  : "Photobucket" 
Bytes Uploaded : 34
Bytes Download : 44
Num Flows : 78

When I run my script I am getting only the last line in $test.

Please help me to fix this.

Thanks,

Kumar

Answer 1

First, I think that you didn't put the regex that you are using in your question because your regex doesn't match at all because of a missing space. It should be:

Photobucket.*(Bytes Uploaded : .* Bytes Download :)

Now, the problem with this regex is that .* is greedy and will match till the end of the string (since it matches anything and everything), and then backtrack one character at a time until the whole regex is matched (that is where the last Bytes Uploaded : and Bytes Download : is matched), or if no match is found, then the regex fails to match. What you need is to make is the .* lazy (or match as little as possible) with the ? modifier:

Photobucket.*?(Bytes Uploaded : .*? Bytes Download :)

The above will match the correct part, except you will have an incorrect value in y since you will also have Bytes Uploaded and such. The trim cannot remove those. You might thus change the regex a bit more:

Photobucket.*?Bytes Uploaded : (\S+):

This will put non space characters, matched by (\\S+) into the variable y . You don't need to trim after that.

---

And you don't need to split and rejoin if you change the regex:

if { [regexp {Photobucket.*?Bytes Uploaded : (\S+)\s} $test - y]} {
    puts "Bytes uploaded : $y"
}

---

To get all the three values, you then just need to add them at the end:

if { [regexp {Photobucket.*?Bytes Uploaded : (\S+)\s+Bytes Download : (\S+)\s+Num Flows : (\S+)\s+} $test - x y z]} {
    puts "Bytes uploaded : $x"
    puts "Byte download : $y"
    puts "Num flows : $z"
}

Answer 2

You can use string commands instead of a giant regex

set stats {"Bytes Uploaded" "Bytes Download" "Num Flows"}
set photobucket_idx [string first Photobucket $test]
foreach stat $stats {
    set digits_start [expr {[string first "$stat : " $test $photobucket_idx] + [string length "$stat : "]}]
    set digits_end [expr {[string first \n $test $digits_start] - 1}]
    set digits($stat) [string range $test $digits_start $digits_end]
}
parray digits

outputs

digits(Bytes Download) = 44
digits(Bytes Uploaded) = 34
digits(Num Flows)      = 78