sync-diff-inspector: change bit_xor to sum

[Leavrth]

What problem does this PR solve?

Issue Number: close #634

What is changed and how it works?

use SUM instead of BIT_XOR.

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has not been approved.

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[GMHDBJD]

Don't think SUM is better then XOR. 🤔

[Leavrth]

Don't think SUM is better then XOR. 🤔

Yes, I think so. But BIT_XOR is easiler to lead to hash collision when the table has no unique column.
I think the difference caused by special failure is easiler to appear than that by the failure such as some bits randomly changed in the data migration. Especially duplicate same rows(a xor a xor a=a, we use count to solve it), a column of some rows is changed in a regular way(#634), ...

So, how about using SUM function only for the table without unique column? @GMHDBJD

[lance6716]

This is not related to UNIQUE column?

[Leavrth]

This is not related to UNIQUE column?

It seems not as simple as I thought. 🥵

[Leavrth]

Change the content of the concat_ws, the results of bit_xor(crc32(...)) of 2 rows from 2 tables are always the same.

mysql> select BIT_XOR(cast(crc32(concat_ws('.', j, i, x, j,x, j)) as unsigned)) as CHECKSUM from db_test.t4;
+------------+
| CHECKSUM   |
+------------+
| 2478727481 |
+------------+
1 row in set (0.01 sec)

mysql> select BIT_XOR(cast(crc32(concat_ws('.', j, i, x, j,x, j)) as unsigned)) as CHECKSUM from db_test.t5;
+------------+
| CHECKSUM   |
+------------+
| 2478727481 |
+------------+
1 row in set (0.00 sec)

mysql> select * from db_test.t5;
+------+-------+------+
| i    | j     | x    |
+------+-------+------+
|    1 | 10001 |    3 |
|    1 | 10011 |    3 |
+------+-------+------+
2 rows in set (0.00 sec)

mysql> select * from db_test.t4;
+------+-------+------+
| i    | j     | x    |
+------+-------+------+
|    1 | 10001 |    2 |
|    1 | 10011 |    2 |
+------+-------+------+
2 rows in set (0.01 sec)

Seems that bit_xor + crc32 is not a good idea. 🤔

[CLAassistant]

All committers have signed the CLA.