Parsing (partially) non-uniform text blocks in Perl - regex

Question

I have a file with a few blocks that look like this in a file (and in a variable, at this point in the program).
Vlan2 is up, line protocol is up
....
reliability 255/255, txload 1/255, rxload 1/255^M
....
Last clearing of "show interface" counters 49w5d
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
....
L3 out Switched: ucast: 17925 pkt, 23810209 bytes mcast: 0 pkt, 0 bytes
33374 packets input, 13154058 bytes, 0 no buffer
Received 926 broadcasts (0 IP multicasts)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
3094286 packets output, 311981311 bytes, 0 underruns
0 output errors, 0 interface resets
0 output buffer failures, 0 output buffers swapped out
Here's a second block, to show you how the blocks can slightly vary:
port-channel86 is down (No operational members)
...
reliability 255/255, txload 1/255, rxload 1/255
...
Last clearing of "show interface" counters 31w2d
...
RX
147636 unicast packets 0 multicast packets 0 broadcast packets
84356 input packets 119954232 bytes
0 jumbo packets 0 storm suppression packets
0 runts 0 giants 0 CRC 0 no buffer
0 input error 0 short frame 0 overrun 0 underrun 0 ignored
0 watchdog 0 bad etype drop 0 bad proto drop 0 if down drop
0 input with dribble 0 input discard
0 Rx pause
TX
147636 unicast packets 0 multicast packets 0 broadcast packets
84356 output packets 119954232 bytes
0 jumbo packets
0 output error 0 collision 0 deferred 0 late collision
0 lost carrier 0 no carrier 0 babble 0 output discard
0 Tx pause
0 interface resets
I want to pick out certain data elements from each block, which may or may not exist in each block. For example, in the first block I posted I may want to know that there are 0 runts, 0 input errors and 0 overrun. In the second block, I might want to know that there are 0 jumbo packets, collisions, etc. If a given query isn't in the block, it's acceptable to just return na, as this is designed to be processed uniformly.
Each block is structured in a similar way to the two I posted; newlines and spaces delimiting some entries, commas delimiting others.
I have a few ideas as to how this might work. I'm unaware if there is any kind of "look back" function in Perl, but I could attempt to look for the field names (runts, "input errors", etc) and then grab the previous integer; that seems like it would be the most elegant solution for this, but I'm unsure if it's possible.
Currently, I'm doing this in Perl. Each "block" that I'm processing is actually several of these blocks (separated by double newlines). It doesn't have to be done in a single regular expressions; I believe it can be done by applying several regular expressions per block. Performance is not really a factor, as this script will run maybe once per hour.
My goal is to get all of this into a .csv file (or some other data format that's easily graphable) in an automated fashion.
Any ideas?
Edit: example output in CSV as I mentioned, which would be written line by line (for multiple entries like this) to a file as the end result. If a particular entry isn't found in the block, it is marked na in the corresponding line:
interface_name,txload,rxload,last_clearing,input_queue,output_drops,runts,....
vlan2,1,1,49w5d,0-75-0-0,0,0,....
port-channel86,1,1,31w2d,na,na,0,...

Answer 1

Simple hash of properties and numbers.
sub extract {
my ($block) = #_;
my %r;
while ($block =~ /(?<num>\d+) \s (?<name>[A-Za-z\s]+)/gmsx) {
my $name = $+{name};
my $num = $+{num};
$name =~ s/\A \s+//msx;
$name =~ s/\s+ \z//msx;
$r{$name} = $num;
}
return %r;
}
my $block = <<'';
Vlan2 is up, line protocol is up
⋮
my $block2 = <<'';
port-channel86 is down (No operational members)
⋮
use Data::Dumper qw(Dumper);
print Dumper {extract $block};
print Dumper {extract $block2};

Answer 2

I don't think a single regex could do it, nor would I want to support it if it could.
Using multiple regexes, you could easily use something like:
(\d+) runts
(\d+) input errors
...etc...
A simple array of property names and a loop could solve this pretty quickly and with very little fuss.
If you can strip down the input to smaller chunks with some preprocessing, you would be less likely to get false positives.

Answer 3

Here is one way to do it in awk, but this needs lots of tweak to be perfect.
But again, use SNMP.
awk '{
printf $1
for (i=1;i<=NF;i++) {
if ($i" "$(i+1)~/Input queue:/) printf ",%s",$(i+2)
if ($i~/runts/) printf ",%s",$(i-1)
if ($i~/multicast,/) printf ",%s",$(i-1)
}
print ""
}' RS="swapped out" file