* a few age related questions
-
- Famous
- Posts: 166
- Joined: 26 Feb 2004 13:07
- Family Historian: V7
- Location: Lincolnshire
- Contact:
Re: a few age related questions
Following on from the above, there appears to be other flaws with dates.
If we use =DayNumber(DateAt(1940,0,0)) I get a value of 708207
Then if we use =CalcDate(DateAt(1,1,1),0,0,708207) I get 22 Dec 1939
Mmmmm
If we use =DayNumber(DateAt(1940,0,0)) I get a value of 708207
Then if we use =CalcDate(DateAt(1,1,1),0,0,708207) I get 22 Dec 1939
Mmmmm
Dave Simpson ~ Boulton, Braham, Carney, Simpson and Jacobs
- tatewise
- Megastar
- Posts: 28341
- Joined: 25 May 2010 11:00
- Family Historian: V7
- Location: Torbay, Devon, UK
- Contact:
Re: a few age related questions
The response from CP is certainly better than I expected.
Your later 'discrepancy' is not a flaw but perfectly correct.
You need to understand the difference between DayNumber and number of Days as explained in the Help pages for the functions.
DayNumber gives a unique number to each day since 1 Jan 0001 taking into account leap years and the changeover date for the Gregorian calendar that skips 10 or 11 calendar dates.
CalcDate takes leap years into account, but it does not allow for the 10 or 11 days that were deducted from the calendar, owing to the changeover date for the Gregorian calendar, when adding or subtracting months and days.
So bearing in mind that 1 Jan 0001 is DayNumber 1 and the 11 days dropped from the calendar:
=CalcDate(DateAt(1,1,1),0,0,708207-1+11) correctly gives 1 Jan 1940
Your later 'discrepancy' is not a flaw but perfectly correct.
You need to understand the difference between DayNumber and number of Days as explained in the Help pages for the functions.
DayNumber gives a unique number to each day since 1 Jan 0001 taking into account leap years and the changeover date for the Gregorian calendar that skips 10 or 11 calendar dates.
CalcDate takes leap years into account, but it does not allow for the 10 or 11 days that were deducted from the calendar, owing to the changeover date for the Gregorian calendar, when adding or subtracting months and days.
So bearing in mind that 1 Jan 0001 is DayNumber 1 and the 11 days dropped from the calendar:
=CalcDate(DateAt(1,1,1),0,0,708207-1+11) correctly gives 1 Jan 1940
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
-
- Famous
- Posts: 166
- Joined: 26 Feb 2004 13:07
- Family Historian: V7
- Location: Lincolnshire
- Contact:
Re: a few age related questions
Thanks Mike, i hadn't thought about leap years and 1752.
Dave Simpson ~ Boulton, Braham, Carney, Simpson and Jacobs
-
- Famous
- Posts: 166
- Joined: 26 Feb 2004 13:07
- Family Historian: V7
- Location: Lincolnshire
- Contact:
Re: a few age related questions
Just heard from CP that in the next release 7.0.9.0 they have fixed 'Estimated Age At'
The “EstimatedAgeAt” built-in function was not returning the min and max values that users might reasonably expect. This has been fixed.
The “EstimatedAgeAt” built-in function was not returning the min and max values that users might reasonably expect. This has been fixed.
Dave Simpson ~ Boulton, Braham, Carney, Simpson and Jacobs
-
- Famous
- Posts: 166
- Joined: 26 Feb 2004 13:07
- Family Historian: V7
- Location: Lincolnshire
- Contact:
Re: a few age related questions
I've just tested the new implementation of =EstimatedAgeAt() and it now gives the expected results.
Prior to this update it returned (for a particular individual) 47 for both MIN and MAX whereas it should have returned 46 for MIN and 47 for MAX.
This is now working correctly
Prior to this update it returned (for a particular individual) 47 for both MIN and MAX whereas it should have returned 46 for MIN and 47 for MAX.
This is now working correctly
Dave Simpson ~ Boulton, Braham, Carney, Simpson and Jacobs
- tatewise
- Megastar
- Posts: 28341
- Joined: 25 May 2010 11:00
- Family Historian: V7
- Location: Torbay, Devon, UK
- Contact:
Re: a few age related questions
I'm sorry to dispel your joy but it still does not work correctly for all types of Dates.
It correctly handles any Birth Range Dates such as btw 1918 and 1922 as it always did.
However, it doesn't handle Death Range Dates such as btw 2018 and 2022 and only detects the first date.
So with those two Date Ranges the Facts tab Age is (95-104) but the EstimatedAgeAt() functions say MIN 95 and MAX 100.
i.e. The function is assuming the Death Date is simply 2018.
Neither the Facts tab nor the EstimatedAgeAt() function take account of before and after dates even though there are younger than < and older than > age qualifiers.
Since you reported the problem to CP initially, it is best if you make them aware of these deficiencies.
It correctly handles any Birth Range Dates such as btw 1918 and 1922 as it always did.
However, it doesn't handle Death Range Dates such as btw 2018 and 2022 and only detects the first date.
So with those two Date Ranges the Facts tab Age is (95-104) but the EstimatedAgeAt() functions say MIN 95 and MAX 100.
i.e. The function is assuming the Death Date is simply 2018.
Neither the Facts tab nor the EstimatedAgeAt() function take account of before and after dates even though there are younger than < and older than > age qualifiers.
Since you reported the problem to CP initially, it is best if you make them aware of these deficiencies.
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
-
- Famous
- Posts: 166
- Joined: 26 Feb 2004 13:07
- Family Historian: V7
- Location: Lincolnshire
- Contact:
Re: a few age related questions
My parade was quite wet enough without you raining on it
I've replied to CP and given them a worked example showing that btw fails on MAX
I've replied to CP and given them a worked example showing that btw fails on MAX
Dave Simpson ~ Boulton, Braham, Carney, Simpson and Jacobs
-
- Famous
- Posts: 166
- Joined: 26 Feb 2004 13:07
- Family Historian: V7
- Location: Lincolnshire
- Contact:
Re: a few age related questions
Latest update from CP
I obviously got a day adrift in the example that I submitted
However, they accept that there is still an issue.
If the earliest possible birth date was 1/1/1918 and the latest possible death date was 31/12/2017, then the maximum possible age would be 99, would it not? I.e. in that scenario, the person would have died one day short of their 100th birthday. However, it still looks wrong. I have logged it for investigation. Thank you for reporting that.
I obviously got a day adrift in the example that I submitted
However, they accept that there is still an issue.
If the earliest possible birth date was 1/1/1918 and the latest possible death date was 31/12/2017, then the maximum possible age would be 99, would it not? I.e. in that scenario, the person would have died one day short of their 100th birthday. However, it still looks wrong. I have logged it for investigation. Thank you for reporting that.
Dave Simpson ~ Boulton, Braham, Carney, Simpson and Jacobs
- tatewise
- Megastar
- Posts: 28341
- Joined: 25 May 2010 11:00
- Family Historian: V7
- Location: Torbay, Devon, UK
- Contact:
Re: a few age related questions
I am not sure that you and CP have fully understood the problem I found.
There may be some minor miscalculations getting one day adrift, but there is a much bigger problem that is years adrift.
With a Death Range Date such as btw 2017 and 2020 the current miscalculations use the earliest date 1/1/2017 and the latest date 31/12/2017, whereas the latest date should be 31/12/2020 three years later.
That miscalculation results in a MAX Age that is 3 years less than it should be and differs from the Facts tab Age.
There may be some minor miscalculations getting one day adrift, but there is a much bigger problem that is years adrift.
With a Death Range Date such as btw 2017 and 2020 the current miscalculations use the earliest date 1/1/2017 and the latest date 31/12/2017, whereas the latest date should be 31/12/2020 three years later.
That miscalculation results in a MAX Age that is 3 years less than it should be and differs from the Facts tab Age.
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry
-
- Famous
- Posts: 166
- Joined: 26 Feb 2004 13:07
- Family Historian: V7
- Location: Lincolnshire
- Contact:
Re: a few age related questions
Hi Mike
The actual example I gave was 1 year out on the MAX, the same principle as the one you gave:
However, if we use a date range for Birth and Death it fails on MAX
e.g.
DOB btw 1918 and 1922 [Earliest date is 01/01/1918; Latest date is 31/12/1922]
DOD btw 2016 and 2017 [Earliest date is 01/01/2016; Latest date is 31/12/2017]
Based on these values we should get the following results:-
=EstimatedAgeAt(%INDI%,%INDI.DEAT.DATE%,MIN,0) Should return 93 [01/01/2016 - 31/12/1922]
The function returns 93 - Correct
However,
=EstimatedAgeAt(%INDI%,%INDI.DEAT.DATE%,MAX,0) Should return 100 [31/12/2017 - 01/01/1918]
The function returns 98 – Incorrect
Unfortunately, in my text comment I said that MAX should return 100 (when I should have said 99).
It's still returning the wrong value! I think a little bit of nit-picking on their behalf?
The actual example I gave was 1 year out on the MAX, the same principle as the one you gave:
However, if we use a date range for Birth and Death it fails on MAX
e.g.
DOB btw 1918 and 1922 [Earliest date is 01/01/1918; Latest date is 31/12/1922]
DOD btw 2016 and 2017 [Earliest date is 01/01/2016; Latest date is 31/12/2017]
Based on these values we should get the following results:-
=EstimatedAgeAt(%INDI%,%INDI.DEAT.DATE%,MIN,0) Should return 93 [01/01/2016 - 31/12/1922]
The function returns 93 - Correct
However,
=EstimatedAgeAt(%INDI%,%INDI.DEAT.DATE%,MAX,0) Should return 100 [31/12/2017 - 01/01/1918]
The function returns 98 – Incorrect
Unfortunately, in my text comment I said that MAX should return 100 (when I should have said 99).
It's still returning the wrong value! I think a little bit of nit-picking on their behalf?
Dave Simpson ~ Boulton, Braham, Carney, Simpson and Jacobs
- tatewise
- Megastar
- Posts: 28341
- Joined: 25 May 2010 11:00
- Family Historian: V7
- Location: Torbay, Devon, UK
- Contact:
Re: a few age related questions
OK, not having seen what you posted, it was difficult to put the CP response in context. I see now.
Mike Tate ~ researching the Tate and Scott family history ~ tatewise ancestry