Okay I took the fastest routine and made it into an $INSERT piece of code:
YEAR = OCONV(THE.DATE, "DY")
IF MOD(YEAR,4) THEN
LEAP.YEAR = 0
END ELSE
IF MOD(YEAR, 100) THEN
LEAP.YEAR = 1
END ELSE
IF MOD(YEAR, 400) THEN LEAP.YEAR = 0 ELSE LEAP.YEAR = 1
END
END
Then created a Function like:
FUNCTION LEAP.YEAR.FUN( THE.DATE )
$INSERT LEAP.YEAR.INS
RETURN LEAP.YEAR
And a Subroutine like:
SUBROUTINE LEAP.YEAR.SUB( LEAP.YEAR, THE.DATE )
$INSERT LEAP.YEAR.INS
RETURN
Then modified Daniel's test program like:
DEFFUN LEAP.YEAR.FUN( THE.DATE )
EQU LC TO 10000000
*
ST = TIME()
FOR THE.DATE = 1 TO LC
LEAP.YEAR = LEAP.YEAR.FUN( THE.DATE )
NEXT THE.DATE
TS = TIME()
CRT TS - ST
*
*
ST = TIME()
FOR THE.DATE = 1 TO LC
CALL LEAP.YEAR.SUB( LEAP.YEAR, THE.DATE )
NEXT THE.DATE
TS = TIME()
CRT TS - ST
*
ST = TIME()
FOR THE.DATE = 1 TO LC
$INSERT LEAP.YEAR.INS
NEXT THE.DATE
TS = TIME()
CRT TS - ST
*
END
The speed results:
Function 24 seconds
Subroutine 21 seconds
Insert 11 seconds
Not surprising results given what we know about these three methods.
Although the function method did get better since the last time I've test
it.
UD 6.1
Other systems will vary.
But this is a good example of working together to enhance the U2 products,
now we all have the same Leap Year calculation and those that were using the
MOD(YEAR, 4) method should now use this method everywhere as to avoid data
corruption.
David A. Green
(480) 813-1725
DAG Consulting
-----Original Message-----
From: Daniel McGrath [mailto:dmcgr...@rocketsoftware.com]
Sent: Wednesday, December 07, 2011 11:57 AM
To: dgr...@dagconsulting.com; U2 Users List
Subject: RE: [U2] End of Month date routine
I added the code to PasteBin so as to not flood here:
http://pastebin.com/JMyqhFud
I did a few more than 100K and I cut the OCONV step out from all of them. I
took the liberty to adjust Will's version to return correct results. The
results I got, in order, was:
16.872
2.804
2.384
5.561
So, if you have a need to account for your milliseconds, the slightly
optimized Wikipedia algorithm seems to be the fastest.
Interested to see if there are any better? (Or if I made any mistakes...)
Regards,
Dan
-----Original Message-----
From: u2-users-boun...@listserver.u2ug.org
[mailto:u2-users-boun...@listserver.u2ug.org] On Behalf Of David A. Green
Sent: Wednesday, December 07, 2011 10:48 AM
To: 'U2 Users List'
Subject: Re: [U2] End of Month date routine
Who wants to take each of the Leap Year calculations and put them into
different subroutines and loop 100K times and see which one is faster using
profiling?
David A. Green
(480) 813-1725
DAG Consulting
_______________________________________________
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users
_______________________________________________
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users
