Can someone explain Math.fmod() vs using a % operator?

[jonathan_s3]

[ FlexSim 20.2.3 ]

Working on a decision point in Flexscript I wanted to use a modulo math operator like I can within Excel. I started tinkering with this:

if (current.labels.assert("roundRobinVal", 1).value % 2) == 0) {...then do stuff...}

But no matter what I did the % did not seem to behave like I anticipated.

I then found Math.fmod() and made this:

if (Math.fmod(current.labels.assert("roundRobinVal", 1).value, 2) == 0) {...then do stuff...}

and it worked perfectly.

I am incrementing the roundRobinVal after the if().

I also don't understand the ", 1" in the latter portion of the .assert().

[jason_lightfoot_adsk]

The 1 is the initial value of the label in the assert -so the value it will take if it didn't exist.

The % operator should work exactly as fmod - can you post an example?

[jonathan_s3]

This works like I intended:

if (Math.fmod(current.labels.assert("roundRobinVal", 1).value, 2) == 0)

{newDest = current.outObjects[1];}

else

{newDest = current.outObjects[2];}

current.labels.assert("roundRobinVal", 1).value++;

This did not:

if ((current.labels.assert("roundRobinVal", 1).value % 2) == 0)

{newDest = current.outObjects[1];}

else

{newDest = current.outObjects[2];}

current.labels.assert("roundRobinVal", 1).value++;

[jason_lightfoot_adsk]

I meant in a model where we can step through and see if something funky is happening, since the description you gave is perfect and should work as far as I can tell.

You could try stepping through with the debugger and put expressions in the watch list to see what everything is evaluating to.

[jonathan_s3]

I've already implemented and saved over using Math.fmod() and it's working now.

What it seemed to be doing is that the even values never evaluated to a 0 remainder with "% 2".

As far as the debugger I haven't used that yet. I'm not sure how to put in breakpoints and watch specific variables in flexsim.

[jason_lightfoot_adsk]

Understood. I really recommend the debugger - it takes out the mystery of many situatons and is pretty simple to use. If you want help on that there's the documentation here or you can reach out for support.

[jonathan_s3]

While I'm thinking about it, can you step me through the logic of:

current.outObjects[(current.labels.assert("roundRobinVal", 1).value++ - 1) % current.outObjects.length + 1];

as in does it evaluate the .value++ and -1 before or after the %.

The way it reads left to right it would read the roundRobinVal (asserting it to 1 if it did not exist already), then increment it by 1, then -1, then divide it by "current.outObjects.length + 1" to get the final value. But that would mean the .value++ -1 would be self defeating so that's no right. I assume the ++ is done after it evaluates the %.

[jason_lightfoot_adsk]

It's incremented after the expression is evaluated. There's no preincrement yet in flexscript afaik.

[Matthew_Gillespie · Accepted solution]

The F in fmod stands for float and doesn't round to the nearest integer like % does, for example:

3.4 % 2                  //    1
Math.fmod(3.4, 2)        //    1.4

[jonathan_s3]

That makes sense with some of the behavior I was seeing. I had been wondering if it was something like that.