grep greater than time period??
 

Hey guys, I'm fairly new at unix shell scripting and I have a quick question.
Quick overview I devolped a script where I generate a file ..and I want to grep any time greater than 30 minutes.
For example i run this which generates a file

./ggsci << endit >$AUDIT_DIR/check_output.out


The file contains:

EMMP 00:00:00 00:00:06
PMMP 00:00:00 00:01:09
RMDRA 00:00:00 00:00:00
RMDRB 00:00:00 00:01:06

I want to grep for any time greater than 30minutes..
How can I do that?

to give a better idea

Program Status Grp lag TimeChkpt

MANAGER

RUNNG
EXTRACT RUN EMMP 00:00:00 00:00:06
EXTRACT RUN PMMP 00:00:00 00:00:06
REPLICAT RUN RMDRA 00:00:00 00:00:00
REPLICAT RUN RMDRB 00:00:00 00:00:06


One column is lag, the other is time check point. If there is a lag and how big of a lag and check point time also...sometimes their is an error and the lag or check point is greater than 30minutes ..

Hope that helps

My thought process is that for a word its something like
grep -w 'Agent is Running and Ready' $AUDIT_DIR/check_2.out

so for time I want it to grep a time greater than 30mintes...

Please explain further, you mean the difference between column 2 and 3 is greater than 30 or the difference between values in a column with relation to the following or preceding line?

what these numbers do is let us know if there is any sort of lag and how much lag. If there is a lag the numbers in column two will increase, if the system has not check pointed for a time frame the numbers in column 3 will increase.

usually the numbers will stay below 5 minutes and if the system is running real slow they might get to 20 minutes.

What i do is runa command to generates the below and puts it into a file:
I run
./ggsci << endit >$AUDIT_DIR/check_oracle_output_$ORACLE_SID.out

the file check_oracle_output_$ORACLE_SID.out contains info below


Program Status Grp lag TimeChkpt

MANAGER

EXTRACT RUN EMMP 00:00:00 00:00:06
EXTRACT RUN PMMP 00:00:00 00:00:06
REPLICAT RUN RMDRA 00:00:00 00:00:00
REPLICAT RUN RMDRB 00:00:00 00:00:06


What I wanted to do was grep the file for anytime greater than 00:30:00 and then i was going to set up an if clause and send myself an e-mail..
hope that helps

Instead of grep you might do this:Basically $target is the max elapse with "!" as a place-holder for the first 4 fields. sort as called above will sort by the 5th field, then sed will delete all lines that come at or before $target. Incidentally, the output will be sorted by the 5th field. The extra echo is so sed works correctly if there isn't anything before $target.
Kevin Barry

Barry thanks for the info. Now this is to check for anything above 30 correct, not just 30 itself?

also for the way it would be is...
your code (target='! ! ! ! 00:00:30'{ echo; echo "$target"; cat 'the-file.txt'; } | sort -t' ' -k5,5 -s | sed '1,/! ! ! !/d') then file name check_oracle_output_$ORACLE_SID.out


like
target='! ! ! ! 00:00:30'{ echo; echo "$target"; cat 'the-file.txt'; } | sort -t' ' -k5,5 -s | sed '1,/! ! ! !/d' check_oracle_output_$ORACLE_SID.out



bare with me, im still a bit new

sorry feel like an idiot, just saw the-file.txt part

Is the code for anything above 30 or just at 30?

I have this
target='! ! ! ! 00:00:30'{ echo; echo "$target"; cat '$AUDIT_DIR/check_oracle_output_$ORACLE_SID.out';} | sort -t' ' -k5,5 -s | sed '
1,/! ! ! !/d'


and i am getting an error on: ./testgg.ksh[31]: syntax error at line 42 : `}' unexpected

Sorry, it needs to be two separate lines. Either that or insert ";" where the newline is. I've used "the-file.txt" in place of the input file and output is to standard output. This should give you everything 30s or greater. If you need the lines in their original order I can help with that, also.
Kevin Barry

Wouldn't that be 30 seconds in the format it is in now? If the lost column is HH:MM:SS (is this correct) are we simply looking for any last column greater than 00:30:00?
If so, how about:
Or if you really want to, skip the output file and just pipe your command into the front.

I was trying to provide a solution that allowed for arbitrary times, e.g. 01:23:45.
Kevin Barry

As am I :)

I am looking for it minutes but i could tell that kevin had it in seconds
Now what im trying to do is basically see if the return status for the above is 0
if it is 0 then echo "time is greater than 30" > $AUDIT_DIR/file1.txt


so in other words could i do
target='! ! ! ! 00:30:00'
{ echo; echo "$target"; cat '$AUDIT_DIR/check_oracle_output_$ORACLE_SID.txt';}| sort -t' ' -k5,5 -s | sed '1,/! ! ! !/d'

if [ $? -eq 0 ]
then
echo "time is greater than 30" > $AUDIT_DIR/file1.txt
fi

So if the target command does succeeds with something above 30 then the text time is greater than 30 would be inputted into the file1.txt...

Am I correct?

Hi,

I've found this with grep only for me it works, with this fileI got this outputwhich as I understood is what you expected.

Markus

You would need to explicitly check for lines being output from sed. You also don't need the if statement.Also, please use [CODE][/CODE] tags for your code or the # button.
Kevin Barry

PS Note that I've used "!" as a placeholder because it has the lowest ASCII of printing characters besides " ", which makes it come out on top of the other lines with a non-empty respective field (for future reference.)