Find out timestamp line and delete if it is older than particular/specific days using bash script
 

Hi All,
I have a file with MAC addresses in which few lines are having epoch value in end of the line and few lines don't have any epoch value. Now I want to grep only those lines which are having epoch value. So could anybody please help me how to grep only epoch value lines?
Please see the below example file.

I take it you just want to extract the lines that have the timestamps? All that takes is an appropriate regex.

(This may even be a bit of overkill, but better safe than sorry.)

I recommend taking the time to learn something about regular expressions. You'll be glad you did.

Here are a few regular expressions tutorials:
http://mywiki.wooledge.org/RegularExpression
http://www.grymoire.com/Unix/Regular.html
http://www.regular-expressions.info/

Invoke:

Respected David the H. Sir, I am very lucky, you came to this thread.
The above code worked successfully.
Actually I want to find out only those MAC address which are having epoch value(The time we added) and added 90 days ago and delete those lines from mac_address file and then reload the squid3 service. This is because whichever MAC address are given in this file will be having internet access for a couple of days/weeks. So we are giving internet access for maximum 90 days after that those device should not access internet. If anybody wanted internet access even after 90 days then we will add those MAC addresses manually with a new timestamp.
But I don't know shell script to achieve this so begging your kind help. So could you please provide a bash script for this i.e searching timestamp, calculating days with current timestamp and if it is reached 90 or more than 90 days then those lines should delete from mac_address file and run the "service squid3 reload" command.
Thanks in advance for your kind help.

Hi shivaa, Very thanks for your kind reply. But it is not extracting whole line instead extracting only end of the field of epoch lines. But I was expecting to extract whole lines as from David The H. command.

Simply do:

Thanks for the reply Shivva,
No, This is extracting all lines which are having more than 4 fields in a line. But if I add something more like below then it extract that line also which I don't want.
But
command can skip such lines even if it is having many fields and print only those lines which are having epoch value.

Hi All,
After reading many threads (especially from David the H.) from linuxquestions.org, I could try for bash script and worked successfully.
For that first I found the epoch value for 90 days using
And according the above I could write a small script as below
To test the above script I have changed timestamp to 90 days back for Anjaneya, Akbar and Suyodhana in mac_address file. Executed script and found working fine by sending mail if it deletes any MAC address.
Hi All, Please let me know if I have done anything wrong in codes.
Once agagin thank you all and Thanks David The H.

If you need to calculate relative values, then you're definitely going to want to use awk instead (or another language like perl that has full mathematic ability).

This will print out every line except ones where the final field is a string of numbers smaller than the epoch - 90 days.

I'll let you flesh out the rest of the script yourself.


Edit the next day: After looking back at this again with fresh eyes, we can simplify the above code even more:

I started by removing the BEGIN section and moved the systime calculation directly into the test. And instead of using next to skip the matching lines, I inverted the test logic with "!" ("not"), so now it will only execute on non-matching lines instead. Finally, since the default action on a true match is to print, we can dump that keyword as well.

Now for a few quick comments on the script you posted.

1)
This section would be more safely and cleanly handled with an array:

In particular parsing ls for filenames or metadata is highly discouraged.

And when using bash or ksh, it's recommended to use ((..)) for numerical tests.

http://mywiki.wooledge.org/ArithmeticExpression

2)
I just demonstrated this above, but in any case this is a useless use of grep. awk is powerful enough to do almost all text processing itself.

It's also not the cleanest way to load an array, since it relies on shell word-splitting. I suggest using mapfile instead, with a process substitution.

3)
This is just a minor point, but unless you need the exact "now" value for each timestamp, it would be more efficient to set the variable once before the loop, rather than resetting it again on every iteration inside it.

4)
"$(..)" is used for command substitution, that is, inserting the output of one command inside another one. The grep command above should not have them.

If you really need to run a command in a subshell, without expansion, use "(..)" brackets instead.

5)
This is using array setting syntax to set line2delete. With the above, "$line2delete" will only hold the first word of the output (it's equal to "${line2delete[0]}").

Assuming the grep output is only one line, use simple command substitution instead.

Finally, this whole script has a rather complex and inefficient code flow, IMO. A better approach would be to simply set the comparison time to a variable first, then loop through the file, using read's ability to load each line into an array, and compare the last field to the timestamp variable.

Or just use awk, as I posted above. ;)

Hi Guru,
I was not aware of your reply. I am really sorry....... for this. And still that script is working fine but still I would like to follow your suggestion.
Thansk a lot.