Fix decode error when "NULL" value in the column with "primary key" flag

[JaySon-Huang]

What problem does this PR solve?

Issue Number: close #5859

Problem Summary:
TiFlash assumes all values in the column with "primary key" flag are non-null value.

However, users could create such "NULL" value in the "primary key" column by following steps

-- create table with `source` is nullable
create table test.t_case (
`case_no` varchar(32) not null,
`source` varchar(20) default null,
`p` varchar(12) DEFAULT NULL,
primary key (`case_no`)
);

-- insert some data and left `source` to be empty
insert into test.t_case(case_no)
values ("1"), ("2"), ("3"), ("4");

-- drop the primary key, fill the `source` to be non-empty
alter table t_case drop primary key;
update test.t_case set `source` = '' where `source` is NULL;
-- add new primary key with case_no and source
alter table t_case add primary key (case_no, source);

-- send the snapshot data to tiflash
alter table t_case set tiflash replica 1;

What is changed and how it works?

• mainly change the logic in addDefaultValueToColumnIfPossible
  • For non-clustered index table(is_common_handle=false && pk_is_handle=false), fill default value even if the column contains primary key flag when the column is not appear in the encoded value(those rows are not supposed to be return to users)
  • For clustered index/pk_is_handle table, ignore the primary key column as they are supposed to be decoded from the encoded key
• Refine the exception message in insertRangeFrom
• Add more region info when exception thrown in preHandleSnapshot/ingestSST

Check List

Tests

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

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

Fix the panic issue after creating the primary index with a column containing `NULL` value

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has been approved by:

• hongyunyan
• lidezhu

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.

[JaySon-Huang]

/run-all-tests

[JaySon-Huang]

/cc @lidezhu @hongyunyan

[JaySon-Huang]

/run-all-tests

[JaySon-Huang]

The helper functions in RegionBlockReaderTestFixture is buggy when testing with primary key columns. I will split it into another PR to refine the test suite first. #5899

[ti-chi-bot]

In response to a cherrypick label: new pull request created: #5934.

[ti-chi-bot]

In response to a cherrypick label: new pull request created: #5935.

[ti-chi-bot]

In response to a cherrypick label: new pull request created: #5936.

[ti-chi-bot]

In response to a cherrypick label: new pull request created: #5937.

[ti-chi-bot]

In response to a cherrypick label: new pull request created: #5938.

[ti-chi-bot]

In response to a cherrypick label: new pull request created: #5939.

[ti-chi-bot]

In response to a cherrypick label: new pull request created: #5940.