Here is another interesting one. To reiterate, these tests were performed by taking two times, calculating the difference between the two times using Application-Business-Time-Diff; then adding that diff to the start time and subtracting that diff from the end time.
* Test 3 Business Time Type: Business Time 1 TestID: 000000000000202 Start: 3/10/2007 11:00:00 PM End: 3/11/2007 11:00:00 PM Diff Seconds: 86400 Diff Minutes: 1440 Diff Hours: 24 Diff Daus: 1 Workday: 24 Hr Tag Holiday: 24 Hr Tag Result of Add to Start (Seconds): 3/11/2007 11:00:00 PM Result of Add to Start (Minutes): 3/11/2007 11:00:00 PM Result of Add to Start (Hours): 3/11/2007 11:00:00 PM Result of Add to Start (Days): 3/11/2007 11:00:00 PM Result of Subtract from End (Seconds): 3/10/2007 11:00:00 PM Result of Subtract from End (Minutes): 3/10/2007 11:00:00 PM Result of Subtract from End (Hours): 3/10/2007 11:00:00 PM Result of Subtract from End (Days): 3/11/2007 12:00:00 AM * DB Values: REQUEST_ID: 000000000000202 STARTDATE: 1173585600 ENDDATE: 1173668400 ENDDATE-STARTDATE: 82800 It appears in this case, that the business time subtract, when using the unit of days, properly accounts for the DST changes (the result has a 1 hr offset from the original time), when all other units do not; at the same time, the business time add of 1 day does not work consistently with subtract with a unit of "days", or I would have seen 3/12/2007 12:00:00 AM as the return value. Notice that the db value for the time difference (end-start) does not match that of the business time calculation (Application-Business-Time-Diff). This is just getting better and better, or should I say the rabbit hole is getting deeper and deeper. Axton Grams On 3/6/07, Axton <[EMAIL PROTECTED]> wrote:
I meant to include this in the original post, but if you look at Test 1, it appears the values are written to the db properly in GMT time, and if I subtract those times (epoch/integers) in the db I get back 3600 (what I expected), but business time 1 gives me a difference of 7200 seconds (2 hours) for the same start/end times. Axton Grams On 3/6/07, Axton <[EMAIL PROTECTED]> wrote: > I am seeing some strange results in my DST testing in regards to business time. > > Environment: > ARServer 7.0.01 patch 001 > SunOS 5.9 Generic_118558-26 > Mid-Tier 7.0.01 patch 001 > IBM WebSphere Application Server/6.0 > IBM Java 1.4.2_11 > > I took 2 dates, a start and end date, performed a diff, then added the > diff to the start and subtracted it from the end; all BT calculations > are done in filters on the server side. These are the results of 3 of > those tests: > > Here are the scenarios I ran using the Mid-Tier (just to avoid > potential client/MS DST issues): > > * Test 0 > Business Time Type: Business Time 1 > TestID: 000000000000109 > Start: 3/10/2007 1:00:00 PM > End: 3/10/2007 2:00:00 PM > Diff Seconds: 3600 > Diff Minutes: 60 > Diff Hours: 1 > Workday: 24 Hr Tag > Holiday: 24 Hr Tag > Result of Add to Start: 3/10/2007 2:00:00 PM (for all 3 units) > Result of Subtract from End: 3/10/2007 1:00:00 PM (for all 3 units) > * DB Values: > REQUEST_ID: 000000000000109 > STARTDATE: 1173549600 > ENDDATE: 1173553200 > ENDDATE-STARTDATE: 3600 > > * Test 1 > TestID: 000000000000111 > Business Time Type: Business Time 1 > Start: 3/11/2007 1:00:00 AM > End: 3/11/2007 3:00:00 AM > Diff Seconds: 7200 > Diff Minutes: 120 > Diff Hours: 2 > Workday: 24 Hr Tag > Holiday: 24 Hr Tag > Result of Add to Start: 3/11/2007 3:00:00 AM (for all 3 units) > Result of Subtract from End: 3/11/2007 1:00:00 AM (for all 3 units) > * DB Values: > REQUEST_ID: 000000000000111 > STARTDATE: 1173592800 > ENDDATE: 1173596400 > ENDDATE-STARTDATE: 3600 > > * Test 2 > TestID: 000000000000115 > Business Time Type: Business Time 1 > Start: 3/11/2007 1:00:00 AM > End: 3/11/2007 2:00:00 AM > Diff Seconds: 3600 > Diff Minutes: 60 > Diff Hours: 1 > Workday: 24 Hr Tag > Holiday: 24 Hr Tag > Result of Add to Start: 3/11/2007 2:00:00 AM (for all 3 units) > Result of Subtract from End: 3/11/2007 1:00:00 AM (for all 3 units) > * DB Values: > REQUEST_ID: 000000000000115 > STARTDATE: 1164780000 > ENDDATE: 1164783600 > ENDDATE-STARTDATE: 3600 > > While the add and subtract put the date back at the expected values, > the diffs seem to be off for test 1 (which covers the DST switch). > > For Test 1, I would expect to see 1 hour as the result of the diff > (1:00-1:59, then jump to 3:00). > > This is very odd. Can someone give me a sanity check on what I think > I should be seeing? The same workflow was used to generate all the > test results (filters on 1 form where I provide the BT version, > holiday/workday/segments, start date, and end date). The calculations > are done properly when using the time difference provided by the diff, > but I am concerned this will cause issues for time calculations where > the offset is provided as a constant (e.g. 3/11/2007 1:00:00 AM + 4 > hours will result in 4/2/2006 5:00:00 AM, when it should result in > 4/2/2006 6:00:00 AM. The 1 hour increments would follow this pattern: > 1-3, 3-4, 4-5, 5-6). > > Thanks, > Axton Grams >
_______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org ARSlist:"Where the Answers Are"