[Babase] Pison date/time notes
Karl O. Pinc
babase@www.eco.princeton.edu
Sat, 9 Oct 2004 02:31:22 -0500
On 2004.10.09 00:52 Susan Alberts wrote:
> Comments below
>
>> This is more of an FYI. Probably because
>> we never did really document the Psion data.
>>
>> 1
>>
>> Note that the times recorded in the current database
>> are truncated to the nearest minute. That is, the
>> current database has two time fields, a datetime
>> value which has full precision, and a time value,
>> which is a string and is truncated to the minute.
>>
>> For instance: INTERACT.Iid # 466615 has a
>> psion time value that's 55 seconds more than
>> the to-the-minute time value. The seconds were
>> lopped off.
>
> This might be a problem if we ever want to look at, for instance,
> time lapsed between two events. When you say "the current database"
> do you mean the foxpro one or the postgresql one or both? Can we
> change this?
By current I mean the foxpro database.
I don't think it'll be a problem.
>>
>> 2
>>
>> In the current database, psion time values for
>> agonisim (and I presume grooming) don't
>> contain the correct date. The date column
>> is correct, and the time value 00:00,
>> but if you look at the date within the Datetime
>> column you get a different date. I suspect
>> that the psion is just using 'day zero'
>> because that has 'time zero' (00:00),
>> which is what's wanted for agonisim and grooming.
>>
>> As the new database has only one datetime column
>> this won't be an issue.
I'm still working my thoughts out regards having
a datetime column or a column for date and a
column for time on point samples. I don't
want to lose the psion data, and it has a
full datetime (I think). I'll work it out.
I'm starting to tend towards having just a
time on POINTS.
Karl <kop@meme.com>
Free Software: "You don't pay back, you pay forward."
-- Robert A. Heinlein