Dates before October 15, 2582
Locked
 
|
When I customize the Begin and End year for dates in the Options window to 1580 to 9900, I receive an error message that the dats must fall betweenn 1582 and 9900. What is the workaround for this limitation?
|
Re: Dates before October 15, 2582
|
|
Ok, this is just too good to pass up.
Please, why do you need to have the date range set to 1580?? I'm assuming you have read the little explanation in the documentation that spss date-time values are stored as the number of seconds since, I think, Oct 15, 1582. I'd bet there is no way around the 1582 date but there are other ways to storing dates that might work for you. Would you describe the problem as it relates to dates that you're working on. Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of spruce18b Sent: Monday, February 28, 2011 3:34 PM To: [hidden email] Subject: Dates before October 15, 2582 When I customize the Begin and End year for dates in the Options window to 1580 to 9900, I receive an error message that the dats must fall betweenn 1582 and 9900. What is the workaround for this limitation? -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Dates-before-October-15-2582-t p3403934p3403934.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Dates before October 15, 2582
|
|
You really cannot go any earlier that October 15, 1582 because on the
recommendation of astronomer and mathematician Christopher Clavius, Pope Gregory XIII removed ten days from the month of October (1582) due the accumulated errors of the Julian calendar which underestimated the length of a year by 11 seconds. SPSS has been keeping close track ever since. Nobody did anything whatsoever between Oct. 4 and Oct. 15, 1582, because the ten intervening days were simply been declared out of existence by the pope. It took a while for the world to accept the change. The British and their colonies held to the Julian calendar until 1752 and the Russians until 1918. JOHN -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Gene Maguin Sent: Monday, February 28, 2011 1:44 PM To: [hidden email] Subject: Re: Dates before October 15, 2582 Ok, this is just too good to pass up. Please, why do you need to have the date range set to 1580?? I'm assuming you have read the little explanation in the documentation that spss date-time values are stored as the number of seconds since, I think, Oct 15, 1582. I'd bet there is no way around the 1582 date but there are other ways to storing dates that might work for you. Would you describe the problem as it relates to dates that you're working on. Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of spruce18b Sent: Monday, February 28, 2011 3:34 PM To: [hidden email] Subject: Dates before October 15, 2582 When I customize the Begin and End year for dates in the Options window to 1580 to 9900, I receive an error message that the dats must fall betweenn 1582 and 9900. What is the workaround for this limitation? -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Dates-before-October-15-2582-t p3403934p3403934.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Dates before October 15, 2582
|
|
In reply to this post by spruce18b
First, the begin and end year specification
in Options only apply to handling of 2-digit year values in date format
variables and is used to determine what century to assign to two-digit
years; so the range cannot exceed 100 years.
Second, if you need to record dates prior to October 15, 1582, you will need to record your dates as three separate standard numeric variables (day, month, year) or as strings. String dates of the format yyyy/mm/dd have a meaningful sort order, but of course you can't calculate differences between string dates. From: spruce18b <[hidden email]> To: [hidden email] Date: 02/28/2011 02:34 PM Subject: Dates before October 15, 2582 Sent by: "SPSSX(r) Discussion" <[hidden email]> When I customize the Begin and End year for dates in the Options window to 1580 to 9900, I receive an error message that the dats must fall betweenn 1582 and 9900. What is the workaround for this limitation? -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Dates-before-October-15-2582-tp3403934p3403934.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
In reply to this post by spruce18b
I know that there were no dates between Oct. 4 and Oct. 15, 1582 and
that there was slow acceptance of the new calendrical regime.
However, when I put the value 14-10-1582 (one day before the change)
into the SPSS data editor with a date format of dd-mmm-yyyy, I
received a missing value and I wondered if there was something wrong
with my understanding of how dates work.
On 2011 February 28 16:36, Rick Oliver wrote: First, the begin and end year specification in Options only apply to handling of 2-digit year values in date format variables and is used to determine what century to assign to two-digit years; so the range cannot exceed 100 years. |
Re: Dates before October 15, 2582
|
|
Indeed, there is something wrong with your understanding. As
mentioned earlier, though you see the date format as dd-mm-yyyy, SPSS stores
the information as a number of seconds since 15-10-1592 00:00:00. This is a limitation of the underlying technique SPSS uses for
storing date information. Jason ___________________________________ Jason Becker Research Specialist Office of Data Analysis and Research Rhode Island Department of Education 255 Westminster Street Providence, RI 02903 (401)-222-8495 From: SPSSX(r) Discussion
[mailto:[hidden email]] On Behalf Of William Klein I know that there were no dates between Oct. 4 and Oct. 15,
1582 and that there was slow acceptance of the new calendrical regime. However,
when I put the value 14-10-1582 (one day before the change) into the SPSS data
editor with a date format of dd-mmm-yyyy, I received a missing value and I
wondered if there was something wrong with my understanding of how dates work. First,
the begin and end year specification in Options only apply to handling of
2-digit year values in date format variables and is used to determine what
century to assign to two-digit years; so the range cannot exceed 100 years.
|
Re: Dates before October 15, 2582
|
|
In reply to this post by spruce18b
Date format values are stored internally
as the number of seconds since the day before October 15, 1582.
From: William Klein <[hidden email]> To: [hidden email] Date: 02/28/2011 10:25 PM Subject: Re: Dates before October 15, 2582 Sent by: "SPSSX(r) Discussion" <[hidden email]> I know that there were no dates between Oct. 4 and Oct. 15, 1582 and that there was slow acceptance of the new calendrical regime. However, when I put the value 14-10-1582 (one day before the change) into the SPSS data editor with a date format of dd-mmm-yyyy, I received a missing value and I wondered if there was something wrong with my understanding of how dates work. On 2011 February 28 16:36, Rick Oliver wrote: First, the begin and end year specification in Options only apply to handling of 2-digit year values in date format variables and is used to determine what century to assign to two-digit years; so the range cannot exceed 100 years. Second, if you need to record dates prior to October 15, 1582, you will need to record your dates as three separate standard numeric variables (day, month, year) or as strings. String dates of the format yyyy/mm/dd have a meaningful sort order, but of course you can't calculate differences between string dates. From: spruce18b <wklein@...> To: [hidden email] Date: 02/28/2011 02:34 PM Subject: Dates before October 15, 2582 Sent by: "SPSSX(r) Discussion" [hidden email] When I customize the Begin and End year for dates in the Options window to 1580 to 9900, I receive an error message that the dats must fall betweenn 1582 and 9900. What is the workaround for this limitation? -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Dates-before-October-15-2582-tp3403934p3403934.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to LISTSERV@... (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Dates before October 15, 2582
Administrator
|
In reply to this post by Jason Becker
Jason meant 1582, not 1592. However, as the following demo shows, datetime variables actually store the number of seconds since 14-Oct-1582 00:00:00. But notice that the minimum valid numeric value is 86400, which is the number of seconds in one day.
data list list / seconds(f12.0). begin data 0 86399 86400 86401 86402 86460 172800 end data. numeric datettimevar (datetime). compute datettimevar = seconds. list. OUTPUT: seconds datettimevar >Warning # 1140 >The value to be output under one of the date formats is negative or has a date >part of zero. The result has been set to the system-missing value. 0 . >Warning # 1140 >The value to be output under one of the date formats is negative or has a date >part of zero. The result has been set to the system-missing value. 86399 . 86400 15-OCT-1582 00:00:00 86401 15-OCT-1582 00:00:01 86402 15-OCT-1582 00:00:02 86460 15-OCT-1582 00:01:00 172800 16-OCT-1582 00:00:00 Number of cases read: 7 Number of cases listed: 7
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
|
|
By way of explanation of the oddities discussed in this thread:
SPSS starts its dates where it does because it calculates days from dates using the algorithm Pope Gregory put in place in 1582. Before that, the calculation was done differently. I think the difference is that Gregory dropped the leap year three centuries out of four, but I could be wrong. SPSS treats values below 86,400 as invalid because they are judged more likely to be mistakes (either days rather than seconds or simply not date values at all) than actual date values. (Suggestion to R&D: instead of displaying the dot, those formats could display the numeric value for those cases.) For working with earlier dates, I'd suggest shifting everything forward a millennium or two. Some calculations will be off by a day. Jonathan Fry On 3/1/2011 9:27 AM, Bruce Weaver wrote: > Jason meant 1582, not 1592. However, as the following demo shows, datetime > variables actually store the number of seconds since 14-Oct-1582 00:00:00. > But notice that the minimum valid numeric value is 86400, which is the > number of seconds in one day. > > data list list / seconds(f12.0). > begin data > 0 > 86399 > 86400 > 86401 > 86402 > 86460 > 172800 > end data. > numeric datettimevar (datetime). > compute datettimevar = seconds. > list. > > OUTPUT: > > > seconds datettimevar > >> Warning # 1140 >> The value to be output under one of the date formats is negative or has a > date >> part of zero. The result has been set to the system-missing value. > 0 . > >> Warning # 1140 >> The value to be output under one of the date formats is negative or has a > date >> part of zero. The result has been set to the system-missing value. > 86399 . > 86400 15-OCT-1582 00:00:00 > 86401 15-OCT-1582 00:00:01 > 86402 15-OCT-1582 00:00:02 > 86460 15-OCT-1582 00:01:00 > 172800 16-OCT-1582 00:00:00 > > Number of cases read: 7 Number of cases listed: 7 > > > > > Jason Becker wrote: >> Indeed, there is something wrong with your understanding. As mentioned >> earlier, though you see the date format as dd-mm-yyyy, SPSS stores the >> information as a number of seconds since 15-10-1592 00:00:00. >> >> >> >> This is a limitation of the underlying technique SPSS uses for storing >> date information. >> >> >> >> Jason >> >> >> >> ___________________________________ >> >> Jason Becker >> >> Research Specialist >> >> Office of Data Analysis and Research >> >> Rhode Island Department of Education >> >> 255 Westminster Street Providence, RI 02903 >> >> (401)-222-8495 >> >> >> >> From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of >> William Klein >> Sent: Monday, February 28, 2011 11:24 PM >> To: [hidden email] >> Subject: Re: Dates before October 15, 2582 >> >> >> >> I know that there were no dates between Oct. 4 and Oct. 15, 1582 and >> that there was slow acceptance of the new calendrical regime. However, >> when I put the value 14-10-1582 (one day before the change) into the >> SPSS data editor with a date format of dd-mmm-yyyy, I received a missing >> value and I wondered if there was something wrong with my understanding >> of how dates work. >> >> On 2011 February 28 16:36, Rick Oliver wrote: >> >> First, the begin and end year specification in Options only apply to >> handling of 2-digit year values in date format variables and is used to >> determine what century to assign to two-digit years; so the range cannot >> exceed 100 years. >> >> Second, if you need to record dates prior to October 15, 1582, you will >> need to record your dates as three separate standard numeric variables >> (day, month, year) or as strings. String dates of the format yyyy/mm/dd >> have a meaningful sort order, but of course you can't calculate >> differences between string dates. >> >> >> >> From: spruce18b<[hidden email]> <mailto:[hidden email]> >> To: [hidden email] >> Date: 02/28/2011 02:34 PM >> Subject: Dates before October 15, 2582 >> Sent by: "SPSSX(r) Discussion"<[hidden email]> >> <mailto:[hidden email]> >> >> ________________________________ >> >> >> >> >> When I customize the Begin and End year for dates in the Options window >> to >> 1580 to 9900, I receive an error message that the dats must fall >> betweenn >> 1582 and 9900. What is the workaround for this limitation? >> >> -- >> View this message in context: >> http://spssx-discussion.1045642.n5.nabble.com/Dates-before-October-15-25 >> 82-tp3403934p3403934.html >> <http://spssx-discussion.1045642.n5.nabble.com/Dates-before-October-15-2 >> 582-tp3403934p3403934.html> >> Sent from the SPSSX Discussion mailing list archive at Nabble.com. >> >> ===================== >> To manage your subscription to SPSSX-L, send a message to >> [hidden email] (not to SPSSX-L), with no body text except the >> command. To leave the list, send the command >> SIGNOFF SPSSX-L >> For a list of commands to manage subscriptions, send the command >> INFO REFCARD >> >> >> >> >> > > ----- > -- > Bruce Weaver > [hidden email] > http://sites.google.com/a/lakeheadu.ca/bweaver/ > > "When all else fails, RTFM." > > NOTE: My Hotmail account is not monitored regularly. > To send me an e-mail, please use the address shown above. > > -- > View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Dates-before-October-15-2582-tp3403934p3405129.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > [hidden email] (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD > > ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Dates before October 15, 1582
|
In reply to this post by spruce18b
At 03:33 PM 2/28/2011, spruce18b wrote:
>When I customize the Begin and End year for dates in the Options window to >1580 to 9900, I receive an error message that the dates must fall between >1582 and 9900. What is the workaround for this limitation? Here's a different point of view: The SPSS date-time representation, which counts seconds starting from the midnight beginning October 15, 1582 as day 1, or (equivalently) second 86,400, can represent earlier times perfectly well, by using smaller values or negative values. What SPSS *can't* do, or anyway has no facility to do, is translate or represent those earlier times in any calendar system. As noted on the list, and in the documentation, SPSS translates to and from only the Gregorian calendar, which did not exist before that starting date. If you need to work with earlier times, you can either choose a different representation (starting point and offset unit), or stay with SPSS's; but in either case, you'll have to write code to convert instants in time from, and to, dates in whatever calendar or calendars you use. I'd consider staying with SPSS's representation, because, a. You'd run less risk of mis-remembering what epoch and offset you ARE using; b. Some SPSS functions -- the TIME.xxx and CTIME.xxx functions -- will still work c. If you're also representing events that fall within the Gregorian calendar, you can use SPSS's formats and translation functions for those, and they will have the correct time interval from pre-Gregorian dates. -Best of luck, Richard Ristow ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |