DSXchange

Hi

Functions like MAX, MIN, CEIL, FLOOR etc work on INT32 datatype.

What if my variables are larger than int32 ?

For example, the signature for MAX is

number1 (int32),number2(int32) returns int32

What if the values i need to compare are int64 ? or even Double. Are there no functions for larger data types than int32 ?

You will get incorrect results when using larger int64 values, when the arguments are expected to be int32. Test it out and see.

Ask IBM Support or the IBM DataStage Product Manager about it. Please let us know if you find anything out.

The Server side of the world has "string" functions to do things like math on large numbers, is there any equivalent here?

Thanks QT_KY. That is precisely what has been happening with us. Overflow and data corruption all over the place. Which is what prompted me to ask this question.

Does anyone know of any options if one needs to use these functions on larger datatypes ?

Chulett - Hi ! Can you please elaborate on that with examples ? Are you saying that there are string functions in server jobs that can be used for mathematical operations of the kind i mentioned in my original post ?

Those functions are documented in the BASIC manual and from what I recall are just for basic math: SADD, SSUB, SMUL and SDIV off the top of my aging head. I was just wondering if there is anything remotely similar on the PX side...

The parallel engine does not have the string math functions to which Craig referred.

And if I remember correctly, the string math functions in the server jobs are VERY slow compared to normal math functions.

Sure... but you only used them on data that the 'normal' functions couldn't handle.