The HCD
The HCD

Reputation: 510

awk with quotes and spaces in bash script

I've the following output in a bash variable set from a received snmp trap:

echo $var

Nov 27 16:20:34 witness logger: trap: vars: DISMAN-EVENT-MIB::sysUpTimeInstance = 0:6:10:29.06,  SNMPv2-MIB::snmpTrapOID.0 = SNMPv2-SMI::enterprises.11.2.29.2.90.0.10000002, SNMPv2 SMI::enterprises.11.2.29.2.90.1 = "Finished Number", SNMPv2-SMI::enterprises.11.2.29.2.90.2 = "Filter Cirteria: [called='3333']", SNMPv2-SMI::enterprises.11.2.29.2.90.3 = "Cleared", SNMPv2     SMI::enterprises.11.2.29.2.90.4 = "major Over-Flow alert on Finished Number for ['3333']", SNMPv2 SMI::enterprises.11.2.29.2.90.5 = "The Corresponding Metric Value is: 0.5", SNMPv2- SMI::enterprises.11.2.29.2.90.6 = "Over-Flow", SNMPv2-SMI::enterprises.11.2.29.2.90.7 = "Tue Nov 27 16:20:05 CET 2012" 

I'm trying to get the following output in variables:

var1 = "Tue Nov 27 16:20:05 CET 2012"
var2 = "Finished Number"
var3 = "The Corresponding Metric Value is: 0.5"
var4 = "Cleared"
var5 = "major Over-Flow alert on Finished Number for ['3333']"

I was thinking of doing this via awk

based on the snmp OID: enterprises.11.2.29.2.90.4, enterprises.11.2.29.2.90.5, 11.2.29.2.90.6 etc...

but can't seem to extract just the content of the quoted content " "

Upvotes: 4

Views: 567

Answers (5)

The HCD
The HCD

Reputation: 510

I ended up using the awk solution, but the others were well suited also. Thanks to all.

val=$(echo $val |  awk '{for(i=1;i<=NF;i++)if($i~/is:/)print $(i+1)}' | cut -d\" -f 1)

For info the script is used in snmptrapd when a trap is received it puts logs in the messages and performs other actions on particular alarms.

The main loop is as follows:

vars=
while read oid val
do
if [ "$vars" = "" ]
  then
    vars="$oid = $val"
  else
    vars="$vars, $oid = $val" 
        if [ "$oid" == "SNMPv2-SMI::enterprises.11.2.29.2.90.5" ]
        then
          val=$(echo $val |  awk '{for(i=1;i<=NF;i++)if($i~/is:/)print $(i+1)}' | cut -d\" -f 1)
          /bin/logger "found: value 5:    $val "
          val5=$val
        fi
 fi
done

Upvotes: 0

Ed Morton
Ed Morton

Reputation: 203254

$ echo "$var" | awk -F\" 'BEGIN{n=split("14 2 10 6 8",v," ")} {for (i=1;i<=n;i++) printf "var%d = \"%s\"\n",i,$(v[i])}'
var1 = "Tue Nov 27 16:20:05 CET 2012"
var2 = "Finished Number"
var3 = "The Corresponding Metric Value is: 0.5"
var4 = "Cleared"
var5 = "major Over-Flow alert on Finished Number for ['3333']"

Also, maybe more what you want, here's how to populate a shell array with the result of the awk execution:

$ IFS=$'\n' varArr=( $(echo "$var" | awk -F\" 'BEGIN{n=split("14 2 10 6 8",v," ")}
 {for (i=0;i<=n;i++) printf "\"%s\"\n",$(v[i])}') )

$ echo "${varArr[1]}"                                                         
"Tue Nov 27 16:20:05 CET 2012"

$ echo "${varArr[2]}"
"Finished Number"

$ echo "${varArr[3]}"
"The Corresponding Metric Value is: 0.5"

$ echo "${varArr[4]}"
"Cleared"

$ echo "${varArr[5]}"
"major Over-Flow alert on Finished Number for ['3333']"

and if you don't want the quotes around your text, just don't add them in the awk script:

IFS=$'\n' varArr=( $(echo "$var" | awk -F\" 'BEGIN{n=split("14 2 10 6 8",v," ")}
{for (i=0;i<=n;i++) print $(v[i])}') )

Both of the above will put the entire input string in ${varArr[0]}. It's a trivial tweak if that's undesirable.

Upvotes: 1

Chris Seymour
Chris Seymour

Reputation: 85775

It seems that you want to match all the strings inside double quotes, which is easiest done with grep:

$ echo $var | grep -o '"[^"]*"'

"Finished Number"
"Filter Cirteria: [called=3333]"
"Cleared"
"major Over-Flow alert on Finished Number for [3333]"
"The Corresponding Metric Value is: 0.5"
"Over-Flow"
"Tue Nov 27 16:20:05 CET 2012"

Explanation:

-o only print the part of the line that matches.

"     # Match opening double quote
[^"]* # Match anything not a double quote
"     # Match closing double quote

Hope this helps you get started.

Upvotes: 3

sampson-chen
sampson-chen

Reputation: 47267

Let's start with something simple so you can see how the fields would be broken down when you use awk:

echo "${var}" | awk 'BEGIN{FS="\""} {for (i=1; i<=NF; i++) {print "["i"]", $i}}'

Or, if your shell supports herestrings:

awk 'BEGIN{FS="\""} {for (i=1; i<=NF; i++) {print "["i"]", $i}}' <<< "${var}"

Output:

[1] Nov 27 16:20:34 witness logger: trap: vars: DISMAN-EVENT-MIB::sysUpTimeInstance = 0:6:10:29.06,  SNMPv2-MIB::snmpTrapOID.0 = SNMPv2-SMI::enterprises.11.2.29.2.90.0.10000002, SNMPv2 SMI::enterprises.11.2.29.2.90.1 = 
[2] Finished Number
[3] , SNMPv2-SMI::enterprises.11.2.29.2.90.2 = 
[4] Filter Cirteria: [called='3333']
[5] , SNMPv2-SMI::enterprises.11.2.29.2.90.3 = 
[6] Cleared
[7] , SNMPv2     SMI::enterprises.11.2.29.2.90.4 = 
[8] major Over-Flow alert on Finished Number for ['3333']
[9] , SNMPv2 SMI::enterprises.11.2.29.2.90.5 = 
[10] The Corresponding Metric Value is: 0.5
[11] , SNMPv2- SMI::enterprises.11.2.29.2.90.6 = 
[12] Over-Flow
[13] , SNMPv2-SMI::enterprises.11.2.29.2.90.7 = 
[14] Tue Nov 27 16:20:05 CET 2012
[15]  

Now choose the fields as required:

var1=$(awk 'BEGIN{FS="\""} {print $14}' <<< "${var}")
var2=$(awk 'BEGIN{FS="\""} {print $2}' <<< "${var}")
var3=$(awk 'BEGIN{FS="\""} {print $10}' <<< "${var}")
var4=$(awk 'BEGIN{FS="\""} {print $6}' <<< "${var}")
var5=$(awk 'BEGIN{FS="\""} {print $8}' <<< "${var}")

Explanation:

  • awk 'BEGIN{FS="\""}: Here we use awk to delimit the input on "
  • {print $14}': Print the specific field enclosed by quotes
  • <<< "${var}": Use herestring instead of echo if available (see above)
  • This is under the assumption that the format of your $var stays relatively consistent with regard to field ordering

Upvotes: 2

choroba
choroba

Reputation: 241808

Perl solution:

echo "$var" | perl -nE 'say "var", ++$x, "=$1" while /(".*?")/g'

Output:

var1="Finished Number"
var2="Filter Cirteria: [called='3333']"
var3="Cleared"
var4="major Over-Flow alert on Finished Number for ['3333']"
var5="The Corresponding Metric Value is: 0.5"
var6="Over-Flow"
var7="Tue Nov 27 16:20:05 CET 2012"

Upvotes: 3

Related Questions