[GLLUG] Is there any sane reason to do this? (Database question.)

[Tim Woodall]

I want to know what was wrong with the fourth and ninth of January
after the millennium.

Given that the table appeared to be populated all the way to 2031, one
would assume there was some sort of algorithm for 'invalid' days.

Tim.


On 08/10/2014, Adrian McMenamin <adrianmcmenamin at gmail.com> wrote:
> On 7 October 2014 20:05, John Winters <john at sinodun.org.uk> wrote:
>
>> I'm extracting data from a legacy (sort of - still actively sold) system
>> which has the following table:
>>
>> (CSV dump of the table)
>>
>>
>> Days,DateIdent
>> 2000-01-03 00:00:00.000,1
>> 2000-01-05 00:00:00.000,2
>> 2000-01-06 00:00:00.000,3
>> 2000-01-07 00:00:00.000,4
>> 2000-01-08 00:00:00.000,5
>> 2000-01-10 00:00:00.000,6
>> 2000-01-11 00:00:00.000,7
>>
>>
> [snip]
>
>
>>
>> I'm no database guru, but this seems to me to be bordering on the
>> insane.  Apart from anything else, it means you have to error check
>> every attempt to retrieve a record with a date in it, in case the date
>> integer does not have a corresponding row in the Date table.
>>
>> Is there a good design reason to do this?
>>
>>
> Somebody didn't want to write the logic to handle dates but just wanted to
> add integers?
>