Invalid floating point operation with round or trunc funct.

migro · Post by **migro** » Tue Jan 01, 2013 8:01 pm

Sorry Marcus,
used the wrong forum. Should have post it into: "Technical / Scripting". Can You please move it?

-----------------------------

When I execute the following (simplified) code, I get an error :"Invalid floating point operation".

Code: Select all

Let>Value=99
Let>Exp=10

Let>floatVar={(%Value%^%Exp%)-1}
mdl>Float: %floatVar%
Let>intVar={round (%floatVar%)}
mdl>Int: %intVar%

With exp=9 (9.13517247483642E17) no error is returned.
Seems that the error occures with numbers >= 1.00000000000000E19

Any idea how to avoid this error?

migro · Post by **migro** » Wed Jan 02, 2013 2:37 am

Think I know now, why an error is thrown here, after reading a little bit more about the round-function.
The floating point value is too big, so it can't be converted to an integer (result will exceed 19 digits...)

Am I right?

Post by **Marcus Tettmar** » Wed Jan 02, 2013 4:52 pm

Your code runs fine for me without error. What am I supposed to be seeing?

Me_again · Post by **Me_again** » Wed Jan 02, 2013 6:32 pm

Could it be a 32 vs 64 bit OS issue? I have the same result as migro on Win7/32, exponent 9 works, exponent 10 results in "Invalid floating point operation".

migro · Post by **migro** » Wed Jan 02, 2013 7:51 pm

Hi together,

I got the error on both XPsp3 (x86) and Win7sp1 Ult (x64).
Please see attached screenshot.

Screenshot was taken on a system with Win7 x64 and 6 GB RAM.

Me_again · Post by **Me_again** » Wed Jan 02, 2013 8:02 pm

So much for my theory

What version of MacroScheduler?

migro · Post by **migro** » Wed Jan 02, 2013 8:15 pm

Version 13.2.4 - November 2012.
I think this was the latest one.

Grovkillen · Post by **Grovkillen** » Thu Feb 07, 2013 9:01 am

Okay, so I ran into this issue myself and when I copy/paste the code from the first post I get these two messages:

First message...

Second message...

EDIT: Version 14.0.2

Grovkillen · Post by **Grovkillen** » Thu Feb 07, 2013 9:19 am

Okay,

I've done some research and the problem is that my system use "," as decimal separator and the round function only work with "." as decimal separator. So I need to convert all the input data to "." from "," which is not very good... is it confirmed that the math functions ony work with "." ?

Grovkillen · Post by **Grovkillen** » Thu Feb 07, 2013 9:26 am

Code: Select all

Let>DECIMAL_SEPARATOR=.
Let>test_function=2.8*1.1
MDL>test_function
Let>what_to_round=1.8
Let>rounded_value={round(%what_to_round%)}
MDL>rounded_value

Let>DECIMAL_SEPARATOR=,
Let>test_function=2,8*1,1
MDL>test_function
Let>what_to_round=1,8
//not working
Let>rounded_value={round(%what_to_round%)}
MDL>rounded_value

Grovkillen · Post by **Grovkillen** » Mon Aug 19, 2013 2:40 pm

Any news regarding this?

Post by **Marcus Tettmar** » Mon Aug 26, 2013 3:24 pm

The math functions in complex expressions (between { and }) only support "." as decimal separator.

Grovkillen · Post by **Grovkillen** » Tue Aug 27, 2013 10:41 am

Yeah I learned this some time ago and even added a comment to the online reference guide (http://www.mjtnet.com/manual/index.html ... ctions.htm). Maybe this should be added to the offline help file as well?

Even better, please make it work with other than "." in the future.

Post by **Marcus Tettmar** » Tue Aug 27, 2013 12:25 pm

We can't make complex expressions work with "," as the decimal separator because it is already the parameter separator. This is common to almost all programming languages. Because "," is the parameter delimiter "." has to be the decimal placeholder.

Outside of Complex Expressions Macro Scheduler obeys regional settings by default, but inside VBScript and Complex Expressions this is not possible.

Grovkillen · Post by **Grovkillen** » Tue Aug 27, 2013 1:25 pm

Thanks Marcus for the information. Future realeases maybe could have a more clear Error message? "Couldn't read string as float" is not enough for me to understand (in the beginning before you enlightened me) that "," was causing the problem...

Invalid floating point operation with round or trunc funct.

Invalid floating point operation with round or trunc funct.

Same here