Round result is different depending on whether its argument is exact or approximate

[zimulala]

Round result is different depending on whether its argument is exact or approximate:
For exact-value numbers, it uses “round half away from zero” rule.
For approximate-value numbers, it uses the “round to nearest even” rule.
See: https://dev.mysql.com/doc/refman/5.7/en/precision-math-rounding.html

1.What did you do?
select cast(25E-1 as signed)

2.What did you expect to see?
+-----------------------+
| cast(25E-1 as signed) |
+-----------------------+
| 2 |
+-----------------------+

3.What did you see instead?
Now they all use “round half away from zero” rule.
+-----------------------+
| cast(25E-1 as signed) |
+-----------------------+
| 3 |
+-----------------------+

[ghost]

Confirming I can still reproduce this against master:

select cast(25E-1 as signed);

..


mysql> select cast(25E-1 as signed);
+-----------------------+
| cast(25E-1 as signed) |
+-----------------------+
|                     3 |
+-----------------------+
1 row in set (0.00 sec)

mysql> select tidb_version()\G
*************************** 1. row ***************************
tidb_version(): Release Version: v4.0.0-beta.2-821-gfe07417e6
Edition: Community
Git Commit Hash: fe07417e6b2f28eb334117d44640d0760d7e9866
Git Branch: master
UTC Build Time: 2020-07-22 12:51:37
GoVersion: go1.13
Race Enabled: false
TiKV Min Version: v3.0.0-60965b006877ca7234adaced7890d7b029ed1306
Check Table Before Drop: false
1 row in set (0.00 sec)

[dyzsr]

The rounding behavior for approximate-value numbers seems to be implementation-defined. Do we have to fix it?

For approximate-value numbers, the result depends on the C library. On many systems, this means that ROUND() uses the “round to nearest even” rule: A value with a fractional part exactly half way between two integers is rounded to the nearest even integer.

[xuhui-lu]

Is this issue still open? If so, I am willing to offer help.

[ichn-hu]

Is this issue still open? If so, I am willing to offer help.

Sure, it is still open, thank you in advance of your fixing PR.

[ichn-hu]

@xuhui-lu you could comment /assign @xuhui-lu to assign yourself on this issue.

[xuhui-lu]

/assign @xuhui-lu

[ti-srebot]

Please edit this comment or add a new comment to complete the following information

Not a bug

1. Remove the 'type/bug' label
2. Add notes to indicate why it is not a bug

Duplicate bug

1. Add the 'type/duplicate' label
2. Add the link to the original bug

Bug

Note: Make Sure that 'component', and 'severity' labels are added
Example for how to fill out the template: #20100

1. Root Cause Analysis (RCA) (optional)

2. Symptom (optional)

3. All Trigger Conditions (optional)

4. Workaround (optional)

5. Affected versions

6. Fixed versions

v5.0.0-rc

[ti-srebot]

( AffectedVersions ) fields are empty.
The values in ( FixedVersions ) fields are incorrect.