Question

I have output from a command in the following format:

Ethernet STATISTICS (ent0) :
Device Type: 2-Port 10/100/1000 Base-TX PCI-X Adapter (14108902)
Hardware Address: 00:09:6b:6e:5d:50
Transmit Statistics:                          Receive Statistics:
--------------------                          -------------------
Packets: 0                                    Packets: 0
Bytes: 0                                      Bytes: 0
Interrupts: 0                                 Interrupts: 0
Transmit Errors: 0                            Receive Errors: 0
Packets Dropped: 0       
ETHERNET STATISTICS (ent1) :
Device Type: 2-Port 10/100/1000 Base-TX PCI-X Adapter (14108902)
Hardware Address: 00:09:6b:6e:5d:50
Transmit Statistics:                          Receive Statistics:
--------------------                          -------------------
Packets: 30                                   Packets: 0
Bytes: 1800                                   Bytes: 0
Interrupts: 0                                 Interrupts: 0
Transmit Errors: 0                            Receive Errors: 0
Packets Dropped: 0                            Packets Dropped: 0
                                              Bad Packets: 0

I need to save to variables the number of packets transmitted associated with ent0 and the number of packets transmitted associated with ent1. I need to use awk for this task and while I know how to extract the number of packets, I don’t know how to associate it with the adapter (ent0 or ent1) listed several lines above it. Seems like I need to use some kind of nested loops but don’t know how to do this in awk.

Answer 1

How about:

# list all ent's and there counts 
$ awk '/ent[0-9]+/{e=$3}/^Packets:/{print e,$2}' file
(ent0) 0
(ent1) 30

# list only the count for a given ent 
$ awk '/ent0/{e=$3}/^Packets:/&&e{print $2;exit}' file
0

$ awk '/ent1/{e=$3}/^Packets:/&&e{print $2;exit}' file
30

Explanation:

The first script prints all ent's along with packets transmitted count:

/ent[0-9]+/        # For lines that contain ent followed by a digit string
{
   e=$3            # Store the 3rd field in variable e
}
/^Packets:/        # Lines that start with Packet:
{
   print e,$2      # Print variable e followed by packet count (2nd field)
}

The second script prints only the count for the given ent:

/ent0/             # For lines that match ent0
{
   e=$3            # Store the 3rd field 
}
/^Packets:/&&e     # If line starts with Packets: and variable e set 
{
   print $2        # Print the packet count (2nd field)
   exit            # Exit the script 
}

You can use command substitution in bash to store the value in shell variable:

$ entcount=$(awk '/ent1/{e=$3}/^Packets:/&&e{print $2;exit}' file)

$ echo $entcount 
30

And the -v option of awk to pass in variables:

$ awk -v var=ent0 '$0~var{e=$3}/^Packets:/&&e{print $2;exit}' file 
0

$ awk -v var=ent1 '$0~var{e=$3}/^Packets:/&&e{print $2;exit}' file 
30