Bug #9706 Precision Math: Sin Function returns incorrect results when used with pi()

• Submitted: 7 Apr 2005 8:28
• Modified: 7 Apr 2005 16:53
• Reporter: Disha
• Status: Not a Bug
• Category: MySQL Server
• Severity: S3 (Non-critical)
• Version: 5.0.3 Beta
• OS: Windows (Windows 2003)
• CPU Architecture: Any

[7 Apr 2005 8:28] Disha

Description:
Using Sin function with any parameter , gives incorrect result

How to repeat:
1. Select Sin(90), Sin(pi(30))//

Actual Result:

+------------------+----------------------+
| sin(90)          | sin(pi())            |
+------------------+----------------------+
| 0.89399666360056 | 1.2246063538224e-016 |
+------------------+----------------------+
1 row in set (0.00 sec)

Expected result:
+------------------+----------------------+
| sin(90)          | sin(pi())            |
+------------------+----------------------+
| 1               | 0            |
+------------------+----------------------+
1 row in set (0.00 sec)

[7 Apr 2005 15:16] Disha

The results are returned in radians. Hence the updated expected results are:

+------------------+----------------------+
| sin(90)          | sin(pi())            |
+------------------+----------------------+
| 0.89399666360056 | 0               |
+------------------+----------------------+
1 row in set (0.00 sec)

[7 Apr 2005 16:53] Jorge del Conde

Hi,

sin() returns a approximate number (double). for approximate numbers, exact comparison does not work. instead of f1 = f2, one should use abs(f1 - f2) < epsilon method of comparison (used in floating point calculations).

Also, sin() is not exact, and pi() is double, so having:

sin(pi()) == 1.2246063538224e-16 is the expected result in this case.

--
Jorge