Prev: Oracle DBA Required Urgently in Blythewood, SC
Next: Quick move tablespace
From: Mladen Gogala on 6 Jan 2010 16:48
On Wed, 06 Jan 2010 21:10:56 +0100, Maxim Demenko wrote:

> Sorry, of course, i am interpreting your words like it better suit my
> needs ;-)
> But either your meant to say "The application has run on 05-Jan-2010" -
> in that case my previous post still applies, or it run indeed on
> 05-Jan-2009 and my math is wrong, but then it is unclear with "The
> result is January 28th, 2009, which is almost a year off".
>
>
> Best regards
>
> Maxim

Maxim, you helped me a lot! It works now. Here is the result, written in
The algorithm is yours, I just had to apply it properly. Thanks again
for your help. I must be getting senile these days.

#!/usr/bin/perl -w
use strict;
sub cnv_tim($);

sub cnv_tim($) {
my $A = shift;
my ( $sec, $min, $hr, $dd, $mm, $yr ) =
localtime( ( $A * 1024 ) / 1000000000 );
$yr += 1900;
$mm++;
my $str = sprintf( "%d-%d-%d %d:%d:%d\n", $yr, $mm, $dd, $hr, $min,
$sec );
return ($str);
}
map { print "tim=$_ --> ", cnv_tim($_); } @ARGV;


Here is how to run it:

mgogala(a)nycwxp2622:/tmp$ ./ttt 1233122413959653
tim=1233122413959653 --> 2010-1-5 13:49:11


You can put as many "tim" timestamps on the command line as necessary,
the script will convert them all, one line per timestamp. Some people
object to my function declarations but I used to program in C and have
carried over some habits.


--
http://mgogala.byethost5.com
From: Steve Howard on 6 Jan 2010 16:51
On Jan 6, 4:46 pm, Steve Howard <stevedhow...(a)gmail.com> wrote:
> On Jan 6, 3:10 pm, Maxim Demenko <mdeme...(a)gmail.com> wrote:
>
>
>
> > On 06.01.2010 21:02, Maxim Demenko wrote:
>
> > > On 06.01.2010 20:27, Mladen Gogala wrote:
> > >> On Wed, 06 Jan 2010 19:37:14 +0100, Maxim Demenko wrote:
>
> > >>> On 06.01.2010 19:02, Mladen Gogala wrote:
> > >>>> On Wed, 06 Jan 2010 17:40:43 +0000, Mladen Gogala wrote:
>
> > >>>>> I am analyzing a trace file and the developer asked me when has this
> > >>>>> event taken place:
>
> > >>>>> =====================
> > >>>>> PARSING IN CURSOR #3 len=159 dep=0 uid=141 oct=3 lid=141
> > >>>>> tim=1233122414291746 hv=3402358638 ad='9a174550' select
> > >>>>> folderarti0_.segment# as col_0_0_, folderarti0_.FOLDER# as col_1_0_
> > >>>>> from FOLDER_ARTICLES folderarti0_ where folderarti0_.segment# in (:1 ,
> > >>>>> :2 , :3 , :4)
> > >>>>> END OF STMT
>
> > >> This is not really helpful. The application has run on 05-Jan-2009. Here
> > >> is what I get when I try using "seconds from 1970":
> > >> The result is January 28th, 2009, which is almost a year off.
>
> > > DATE'1970-01-01'+(1
> > > -------------------
> > > 05.01.2010 18:49:12
>
> > > Best regards
>
> > > Maxim
>
> > Sorry, of course, i am interpreting your words like it better suit my
> > needs ;-)
> > But either your meant to say "The application has run on 05-Jan-2010" -
> > in that case my previous post still applies, or it run indeed on
> > 05-Jan-2009 and my math is wrong, but then it is unclear with "The
> > result is January 28th, 2009, which is almost a year off".
>
> > Best regards
>
> > Maxim
>
> It doesn't work for me either...
>
> SQL> !date
> Wed Jan  6 16:44:47 EST 2010
>
> SQL> alter session set events '10046 trace name context forever, level
> 1';
>
> Session altered.
>
> SQL> select 1 from dual;
>
>          1
> ----------
>          1
>
> SQL> !ls -lrt | tail -1
> -rw-r----- 1 oracle dba   1923 2010-01-06 16:45 wcasprod_ora_3974.trc
>
> SQL> !grep PARSE wcasprod_ora_3974.trc
> PARSE
> #3:c=0,e=34,p=0,cr=0,cu=0,mis=0,r=0,dep=0,og=1,tim=1233217089752915
> PARSE
> #2:c=0,e=1029,p=0,cr=0,cu=0,mis=1,r=0,dep=0,og=1,tim=1233217094339231
> PARSE
> #3:c=0,e=39,p=0,cr=0,cu=0,mis=0,r=0,dep=0,og=1,tim=1233217094340214
>
> SQL> select to_date('1970-01-01','YYYY-MM-DD') + (&1/1000000) / 86400
> from dual
>   2  /
> Enter value for 1: 1233217089752915
>
> TO_DATE('1970-01-01','YYYY-MM
> -----------------------------
> 29-JAN-2009 08:18:10
>
> SQL>

I think I found the discrepancy, the 1024 mackes the difference.

If this is true, then the time is not truly quoted in microseconds (as
Oracle documents it) in the trace file?
From: Maxim Demenko on 6 Jan 2010 17:22
On 06.01.2010 22:48, Mladen Gogala wrote:
> On Wed, 06 Jan 2010 21:10:56 +0100, Maxim Demenko wrote:
>
>> Sorry, of course, i am interpreting your words like it better suit my
>> needs ;-)
>> But either your meant to say "The application has run on 05-Jan-2010" -
>> in that case my previous post still applies, or it run indeed on
>> 05-Jan-2009 and my math is wrong, but then it is unclear with "The
>> result is January 28th, 2009, which is almost a year off".
>>
>>
>> Best regards
>>
>> Maxim
>
> Maxim, you helped me a lot! It works now. Here is the result, written in
> The algorithm is yours, I just had to apply it properly. Thanks again
> for your help. I must be getting senile these days.
>
> #!/usr/bin/perl -w
> use strict;
> sub cnv_tim($);
>
> sub cnv_tim($) {
> my $A = shift;
> my ( $sec, $min, $hr, $dd, $mm, $yr ) =
> localtime( ( $A * 1024 ) / 1000000000 );
> $yr += 1900;
> $mm++;
> my $str = sprintf( "%d-%d-%d %d:%d:%d\n", $yr, $mm, $dd, $hr, $min,
> $sec );
> return ($str);
> }
> map { print "tim=$_ --> ", cnv_tim($_); } @ARGV;
>
>
> Here is how to run it:
>
> mgogala(a)nycwxp2622:/tmp$ ./ttt 1233122413959653
> tim=1233122413959653 --> 2010-1-5 13:49:11
>
>
> You can put as many "tim" timestamps on the command line as necessary,
> the script will convert them all, one line per timestamp. Some people
> object to my function declarations but I used to program in C and have
> carried over some habits.
>
>

I'm glad, it works for you.
The credit goes of course to Andrew Zitelli, i didn't participate in his
test, but tested his assumption a couple of times on systems i had
access to - and it seems to be the correct rule.

@Steve
i agree, it seems, the official documentation is incomplete, keep in
mind the differences between solaris/linux and windows (and ...) platforms.

Best regards

Maxim
From: Steve Howard on 6 Jan 2010 18:15
On Jan 6, 5:22 pm, Maxim Demenko <mdeme...(a)gmail.com> wrote:
> On 06.01.2010 22:48, Mladen Gogala wrote:
>
>
>
>
>
> > On Wed, 06 Jan 2010 21:10:56 +0100, Maxim Demenko wrote:
>
> >> Sorry, of course, i am interpreting your words like it better suit my
> >> needs ;-)
> >> But either your meant to say "The application has run on 05-Jan-2010" -
> >> in that case my previous post still applies, or it run indeed on
> >> 05-Jan-2009 and my math is wrong, but then it is unclear with "The
> >> result is January 28th, 2009, which is almost a year off".
>
> >> Best regards
>
> >> Maxim
>
> > Maxim, you helped me a lot! It works now. Here is the result, written in
> > The algorithm  is yours, I just had to apply it properly. Thanks again
> > for your help. I must be getting senile these days.
>
> > #!/usr/bin/perl -w
> > use strict;
> > sub cnv_tim($);
>
> > sub cnv_tim($) {
> >      my $A = shift;
> >      my ( $sec, $min, $hr, $dd, $mm, $yr ) =
> >        localtime( ( $A * 1024 ) / 1000000000 );
> >      $yr += 1900;
> >      $mm++;
> >      my $str = sprintf( "%d-%d-%d %d:%d:%d\n", $yr, $mm, $dd, $hr, $min,
> > $sec );
> >      return ($str);
> > }
> > map { print "tim=$_ -->  ", cnv_tim($_); } @ARGV;
>
> > Here is how to run it:
>
> > mgogala(a)nycwxp2622:/tmp$ ./ttt 1233122413959653
> > tim=1233122413959653 -->  2010-1-5 13:49:11
>
> > You can put as many "tim" timestamps on the command line as necessary,
> > the script will convert them all, one line per timestamp. Some people
> > object to my function declarations but I used to program in C and have
> > carried over some habits.
>
> I'm glad, it works for you.
> The credit goes of course to Andrew Zitelli, i didn't participate in his
> test, but tested his assumption a couple of times on systems i had
> access to - and it seems to be the correct rule.
>
> @Steve
> i agree, it seems, the official documentation is incomplete, keep in
> mind the differences between solaris/linux and windows (and ...) platforms.
>
> Best regards
>
> Maxim

This is the best news I have heard this week. Thanks!
From: joel garry on 6 Jan 2010 18:22
On Jan 6, 2:22 pm, Maxim Demenko <mdeme...(a)gmail.com> wrote:
> On 06.01.2010 22:48, Mladen Gogala wrote:
>
>
>
> > On Wed, 06 Jan 2010 21:10:56 +0100, Maxim Demenko wrote:
>
> >> Sorry, of course, i am interpreting your words like it better suit my
> >> needs ;-)
> >> But either your meant to say "The application has run on 05-Jan-2010" -
> >> in that case my previous post still applies, or it run indeed on
> >> 05-Jan-2009 and my math is wrong, but then it is unclear with "The
> >> result is January 28th, 2009, which is almost a year off".
>
> >> Best regards
>
> >> Maxim
>
> > Maxim, you helped me a lot! It works now. Here is the result, written in
> > The algorithm  is yours, I just had to apply it properly. Thanks again
> > for your help. I must be getting senile these days.
>
> > #!/usr/bin/perl -w
> > use strict;
> > sub cnv_tim($);
>
> > sub cnv_tim($) {
> >      my $A = shift;
> >      my ( $sec, $min, $hr, $dd, $mm, $yr ) =
> >        localtime( ( $A * 1024 ) / 1000000000 );
> >      $yr += 1900;
> >      $mm++;
> >      my $str = sprintf( "%d-%d-%d %d:%d:%d\n", $yr, $mm, $dd, $hr, $min,
> > $sec );
> >      return ($str);
> > }
> > map { print "tim=$_ -->  ", cnv_tim($_); } @ARGV;
>
> > Here is how to run it:
>
> > mgogala(a)nycwxp2622:/tmp$ ./ttt 1233122413959653
> > tim=1233122413959653 -->  2010-1-5 13:49:11
>
> > You can put as many "tim" timestamps on the command line as necessary,
> > the script will convert them all, one line per timestamp. Some people
> > object to my function declarations but I used to program in C and have
> > carried over some habits.
>
> I'm glad, it works for you.
> The credit goes of course to Andrew Zitelli, i didn't participate in his
> test, but tested his assumption a couple of times on systems i had
> access to - and it seems to be the correct rule.

I did participate, but totally forgot until you posted this, though it
sounded familiar... :-)

>
> @Steve
> i agree, it seems, the official documentation is incomplete, keep in
> mind the differences between solaris/linux and windows (and ...) platforms.
>

Not to mention changes and bugs over time, so to speak.

jg
--
@home.com is bogus.
"• Prevent privileged IT staff from accessing the data they are
charged with managing" - Roxana Bradescu Senior Director, Oracle
Database Security Product Marketing


First  |  Prev  | 
Pages: 1 2 3
Prev: Oracle DBA Required Urgently in Blythewood, SC
Next: Quick move tablespace