In our app, we do a lot of date arithmetic including day of week calculations. In AREV we would calculate day of week as follows:

  DOW=MOD(Desired_Date,7)+1 

which returns a day number from 1 to 7 with day 1 being Sunday. In OI if you do this on a date that has a negative sign in internal format, you get a negative number. bummer

The only way around it is to compute MOD yourself:

  DOW=Desired_Date - INT((Desired_Date/7)*7)+1

Anybody else found this???

Don Miller

C3 Inc.

Don,

It sounds familiar. I think I must have come across it. I even thought it was documented somehow.

But it isn't. Actually the docs say:

The modulo (remainder) is calculated by the following formula: Mod(X,Y)=X - (Int(X/Y) * Y)

Which is a lie.

But it also hints at a workaround: Use the formula

(Int(X/Y) * Y)

instead!

I'm certain somebody else must have seen it, e.g when trying to calculate day of week before day 0 according to Revelation.

- Oystein -

Without actually testing this, how about…

DOW=ABS(MOD(Desired_Date,7))+1

Aaarg! Here I am with my degree in Maths and manage to get it wrong! My excuse is I haven't got a degree in Cut & Paste.

Here - with some luck - is the correct version:

The docs say about the Mod function that it is

calculated by the following formula: Mod(X,Y)=X - (Int(X/Y) * Y)

This formula is correct and works for all integers - also when X is negative. So one can use that instead of the Mod function.

OI's Mod fails miserably on negative numbers, so it's a lie that the Mod funcion calculates according to that formula.

To get DOW set Y=7 and add 1:

DOW(X)=X - (Int(X/7) * 7) + 1

(The range of a properly working Mod(X, Y) function is {0, 1, 2, …, Y-1}, while the range of DOW is shifted 1 higher.)

Mike's and Paul's formulas are both wrong. They fail on negative numbers.

Here's a function that compares all four of them - OI's Mod vs the correct formula vs Mike's formula vs Paul's formula. (I've subtracted 1 from Mike's and Paul's formulas since their formulas are for DOW and not Mod):

function Compare_Mod( N, M )
return "OI Mod " : Mod(N, M) : " vs correct formula " : N - (Int(N/M) * M) : " vs Mike " : Abs(Mod(N, M) + 1) - 1 : " vs Paul " : Abs(Mod(N, M)) 

To see the result on -10 function Test_Mod( Dummy )

  equate CrLf to Char(13) : Char(10)
  declare function Compare_Mod
  List="
  for N=-10 to 10
      List := N : " mod " : 7 : ": " : Compare_Mod( N, 7 ) : CrLf
  next N

return List

</code>

- Oystein -

I dont think its a bug. The MOD function is suppose to supply the remainder. And well the remainder is negative for a negative division.

Just to let you know… That is how the MOD function works for many different basic programming applications.

Richard,

Did somebody say it's different in Arev and OI? Then it's a bug.

And Mod not working according to the docs means it's either a bug or an error in the docs.

Finally - if the OI Mod function is supposed to be a mathematical modulo function it's a bug. The mathematical modulo function sort of wraps the number line around a circle. E.g the positive end clockwise and the negative end counterclockwise. Like a time line around a watch face. So the time at -13 hours relative zero (midnight or midday) is not -1 (the remainder -1 after dividing by 12 (the circumference)). It's 11.

- Oystein -

View this thread on the forum...