String manipulation for a date

[joeferns79]

A file contains the following contents ...

20090224T105933 s990atap01      Statistics      0       curam.util.internal.workflow.intf.WorkflowQueueHandler.handleWorkflow
EnactmentMessage ( java.lang.String java.lang.String java.lang.Long )   true    5608
20090224T105940 s990atap01      Statistics      0       curam.util.internal.workflow.intf.WorkflowActivityQueueHandler.handle
ActivityMessage ( java.lang.String java.lang.Long java.lang.Long java.lang.Long java.lang.String )      true    7308

I want to replace the date/time string with the following....

2009-02-24-10.59.33.000000 ......
2009-02-24-10.59.40.000000 .....

Of course, this has to repeat for the complete file.

Is there a Perl command that would do it?

[KevinADC]

There is no perl command to do that but you could use a regexp to capture the date part of the string/line (or maybe another way depending on how the actual data is) and use unpack (or a regexp) and sprintf to reformat it like you want.

[joeferns79]

What would the regular expression look like?

[KevinADC]

What would the regular expression look like?

probably something like:

/searchpattern/

[joeferns79]

Man, you truly are an Expert, aren't you? :-)

[KevinADC]

My wife always says I am. My kids are generally less complimentary though.

[joeferns79]

Kevin,

I got the following code for this ...

#!/usr/bin/perl

$mystring = "20090216T131131";
print "The input is $mystring";
printf "\n";
@myarray = ($mystring =~ m/(\d+)/g);
$result = join("", @myarray);
my ($y,$m,$d,$h,$min,$sec) = unpack("A4A2A2Z2Z2Z2",$result);
$done = sprintf ("%d-%2.2d-%d-%d.%d.%d.%06s",$y,$m,$d,$h,$min,$sec,0);
print "$done";
printf "\n";

which returns the following ...

bash-3.00$ ./test.pl

The input is 20090216T131131
2009-02-16-13.11.31.000000

The problem is i've got a file that contains a number of occurrences of the timestamp, eg

20090224T105710 fem.facade.intf.caseSearch
20090224T105710 codetable.intf.CodeTable
20090224T105719 intf.FieldLevelSecurityRuntime
20090224T105723 resolveCase
20090224T105724 FieldLevelSecurity

If I need to use the code above to change all the occurences of a timestamp, one at a time, to the format specified, and feed those back into the file, how should I go about it?

i.e. I want to change 20090224T105710 to 2009-02-24-10.57.10.000000 ,
20090224T105719 to 2009-02-24-10.57.19.000000 , and so on.

[KevinADC]

open(my $data, "<", "path/to/your/file") or die "$!"; 
while(my $mystring = <$data>){
   print "The input is $mystring";
   my ($date, $otherstuff) = split(/\s+/,$mystring);
   my ($y,$m,$d,$h,$min,$sec) = unpack("A4A2A2Z2Z2Z2",$date);
   my $done = sprintf ("%d-%2.2d-%d-%d.%d.%d.%06s",$y,$m,$d,$h,$min,$sec,0);
   print "$done\n";
}
close $data;

[numberwhun]

Please use square brackets for code tags instead of angle brackets and you will get the desired effect that we require. Thanks!

Regards,

Jeff

[joeferns79]

Kevin,

After changing my code I now get only the dates ...

2009-02-24-10.57.10.000000
2009-02-24-10.57.10.000000
2009-02-24-10.57.19.000000
2009-02-24-10.57.23.000000
2009-02-24-10.57.24.000000
2009-02-24-10.57.50.000000
2009-02-24-10.57.50.000000
2009-02-24-10.57.50.000000
2009-02-24-10.57.50.000000
2009-02-24-10.57.50.000000
2009-02-24-10.51.8.000000
2009-02-24-10.51.8.000000
2009-02-24-10.51.8.000000
2009-02-24-10.51.16.000000
2009-02-24-10.54.4.000000
2009-02-24-10.54.4.000000

I do not get the rest of the line, i.e.

I should be getting ...

2009-02-24-10.57.10.000000 fem.facade.intf .caseSearch
2009-02-24-10.57.10.000000 codetable.intf. CodeTable
2009-02-24-10.57.19.000000 intf.FieldLevel SecurityRuntime
2009-02-24-10.57.23.000000 resolveCase

.. so on

[KevinADC]

change:

 print "$done\n";

to:

   print "$done $otherstuff\n";

[joeferns79]

Thanks, Kevin. Appreciate it.

[markhu]

IMHO, unpack seems like a lot of extra work.

Try this one-liner using regular expression:

$ perl -pi.bak -e 's/(....)(..)(..)T(..)(..)(..)/$1-$2-$3-$4.$5.$6.000000/' /tmp/junk.txt

[numberwhun]

IMHO, unpack seems like a lot of extra work.

Try this one-liner using regular expression:

$ perl -pi.bak -e 's/(....)(..)(..)T(..)(..)(..)/$1-$2-$3-$4.$5.$6.000000/' /tmp/junk.txt

And if your going to do that, take it one step further and do:

$ perl -pi.bak -e 's/(\d{,4})(\d{,2})(\d{2})T(\d{,2})(\d{,2})(\d{,2})/$1-$2-$3-$4.$5.$6.000000/' /tmp/junk.txt

A bit more elegant, but also untested at the moment, but should work none the less (unless I, without looking, messed up the options in the curly brackets).

By the way, please do use code tags around any code you put into the forums. Even command likes as you have put here as they essentially have code in them.

Regards,

Jeff

[nithinpes]

Jeff's command should have been:

perl -pi.bak -e 's/(\d{4})(\d{2})(\d{2})T(\d{2})(\d{2})(\d{2})/$1-$2-$3-$4.$5.$6.000000/' /tmp/junk.txt

\d{,4} will set an upper limit of 4 occurences for digits(max. 4 digits) and not exact 4 occurences as in \d{4} (which is appropriate for the given sample data).