Need pattern matching commands from a log file

[Shivkumar]

Hello,

I am looking out for a pattern in a log file say somefile.log

The log gets entries like below sometime.

Success (code 0) :10001

Success (code 0) :9104

Success (code 0) :9021

I need to grep string "Success (code 0)" and number if it is less than 9000 in the log file.

once the above pattern match is successful then i will need to assign it to a variable called TEMPFILE like below:

TEMPFILE= `Pattern matching commands`

Appreciate your help.

Thanks,
Shiv

[mvpel]

Do you have codes lower than 1000? If you can assume it's always going to be four digits it makes the grep pattern easier.

[James R. Ferguson]

Hi Shiv:

# VALS=$(perl -nle 'm{Success.+\s:(\d+)} && $1 > 9000 and print $1' file

Regards!

...JRF...

[James R. Ferguson · Accepted solution]

Hi (again):

Oops: you wanted less than 9000:

# VALS=$(perl -nle 'm{Success.+\s:(\d+)} && $1 < 9000 and print $1' file

Regards!

...JRF...

[Shivkumar]

Is it possible to have shell script commands as well ?

[Steven E. Protter]

Shalom Shiv,

That last perl command looks pretty good to me.

One command line that does it all.

Partial attempt below...

I could do it multi step in the command line.

grep Success > tfile1

while read -r DL
do
CODE=$(echo $DL | awk -F: '{ print $2}')
< perhaps a case statement here >

done < tfile1

SEP

[James R. Ferguson]

Hi (again) Shif:

> Is it possible to have shell script commands as well ?

Given your format, something like this would work for you:

# cat ./findit
#!/usr/bin/sh
while read LINE
do
echo "${LINE}" | grep -q Success || continue
VAL=$( echo "${LINE}" | cut -d: -f2 )
if [ "${VAL}" -lt 9000 ]; then
VALUES=$( echo "${VALUES} ${VAL}" )
fi
done < $1
echo ${VALUES}

...run as:

# ./findit file

Regards!

...JRF...

[Shivkumar]

Forgot to mention that log file will contain some other lines containing different strings for errors.

Some of the lines will be as mentioned below:
Success (code 0) :10001

Success (code 0) :12004

Success (code 0) :9021

[Patrick Wallek]

That looks the same as what you originally posted.

[James R. Ferguson]

Hi (again) Shiv:

> Forgot to mention that log file will contain some other lines containing different strings for errors.

The code (both Perl and shell) I posted reports the values that follow a ":". If the line doesn't contain the string "Success" then the line isn't a candidate and is skipped.

Regards!

...JRF...

[Steven E. Protter]

Shalom again Shiv,

Take note:

VAL=$( echo "${LINE}" | cut -d: -f2 )

That line of excellent shell code assumes the data we need to look at comes after the :

If that is not the case with all your data, further adjustment will be required.

SEP

[James R. Ferguson]

Hi:

> SEP: Take note...That line of excellent shell code assumes the data we need to look at comes after the :

Yes, that's what I said!

> SEP: If that is not the case with all your data, further adjustment will be required.

That goes without saying. After all, that's what good parsing is all about.

...JRF...

[Dennis Handly]

>once the above pattern match is successful then I will need to assign it to a variable called TEMPFILE like below:
TEMPFILE= `Pattern matching commands`

You want TEMPFILE to have a copy of ALL lines matching your pattern? Or only the first?

TEMPFILE=$(awk -F: '
/Success \(code 0\)/ {
if ($2 < 9000) {
print $0
exit # if want only first
}
}' file)

[Shivkumar]

Dennis,

I put your code in an script and ran it in debug mode by putting first line as #!/bin/ksh -x).

Got below output:

+ + awk -F:
/Success \(code 0\)/ {
if ($2 < 9000) {
print $0
exit # if want only first
}
} bh.log
TEMPFILE=Success (code 0) :10002

============================
Looks like it is wrongly comparing.

[Dennis Handly]

>Looks like it is wrongly comparing.

I can't get it to fail like you have. It only fails for me if you have something after the number:
Success (code 0) :80210 xx

I fixed it with: if ($2 + 0 < 9000) {

[Arturo Galbiati]

Hello,
try:
TEMPFILE=$(awk -F: '
/Success \(code 0\):/ {
if ($2 < 9000) {
print $0
exit # if want only first
}
}' file)

Note ":" added. In this way in the second file ($2( you shoulkd have only numeric value.

HTH,
Art

[Dennis Handly]

>Arturo: /Success \(code 0\):/ {
>Note ":" added. In this way in the second field ($2) you should have only numeric value.

This doesn't add/fix anything unless you want to make sure that ":" is matched.

[Arturo Galbiati]

Hi Dennis,
you're rigth!
Sorry I dind't read comamnd carefully: -F: was already at the beginning.
Discard my reply!
Rgds,
Art