Data type conversion from P(decimal) to type F(float)

[former_member312712]

Hi,

I have following problem:

DATA:
  l_value_p            TYPE p LENGTH 11 DECIMALS 8 value '33.33333333',
  l_value_f            TYPE f.

l_value_f = l_value_p.

The result is that l_value_f has value 3.3333333330000002E01. How come there is 2 at the end? I need value 3.3333333330000000E01. Seems to be some rounding issue. Does anybody know how to get rid of this rounding?

Thanks in advance,

Tomas

[Former Member]

Hi,

The number of bytes in both cases do not match correctly.

Hence i believe a garbage value is inserted.

try the following code.

DATA:

l_value_p TYPE p LENGTH 11 DECIMALS 7 value '33.33333333', " value

l_value_f TYPE f.

l_value_f = l_value_p.

Write: l_value_f.

hope this helps,

Raj

[former_member312712]

Hi Raj,

thanks for your solution, in this way, the value is really converted precisely, but I simply cannot cut the last decimal number, I have to consider all 8 decimal numbers.

Any other idea?

[Former Member]

Hi Tomas,

Try this piece of code.

DATA:

l_value_p TYPE p LENGTH 11 DECIMALS 7 value '33.33333333', " value

l_char(11),

l_value_f TYPE f.

l_char = l_value_p.

l_value_f = l_char.

Write: l_value_f.

-Regards,

Raj

Edited by: Rajat Chaturvedi on Nov 27, 2008 5:30 PM

[former_member312712]

Hi Raj, this solution with l_char does not work for number with 8 decimal places either.

[Former Member]

Hi,

Execute this code.

DATA:

l_value_p TYPE p LENGTH 11 DECIMALS 8 value '33.33333333',

l_value_f TYPE float.

(I just changed type f to float)

l_value_f = l_value_p.

write l_value_f.

OUTPUT: 3.333333333000000E+01

Thanks

Nitesh

[former_member312712]

Hi Nitesh,

your solution surprised me - just changing "F" to "FLOAT".

Unfortunately, it does not work either. The value is printed correctly by WRITE but internally inside variable l_value_f it still holds value "3.3333333330000002E+01" and this is my problem.

I need the internal value of l_value_f to be 3.3333333330000000E+01.

To my understanding, this is problem cannot be solved.

This is because of type float - the conversion of any kind of numbers into float number is very dangerous and the result is unpredictable. I think we have to reconcider the usage of float in our application art the first place.

Thanks anyway, Tomas

[Former Member]

Hi,

Did you check my response?

Thanks

[Former Member]

Tomas,

It works!!!

I can see the result on my system.

Can you plz check again and let me know.

Regards,

Raj

[former_member312712]

Hi Raj, you are right, it works but only if the DECIMAL is 7, not if the DECIMAL is 8. And I cannot use DECIMAL 7 because I must not cut the last number.

That means, I need value

33,33333333

and not value

33,3333333

Thanks anyway, Tomas

[former_member312712]

As mentioned above, there is no solution for this because float is never so precise. Not all number can be described by float, for example this number.