Add random name suffix to SQL Server test tables

[LBPeraza]

Description

Multiple test functions create tables with names from testTableNameTestData, which can cause flaky failures when running them concurrently. Adding these random suffixes should make sure the table names never collide.

Example stack trace:

io.trino.testing.QueryFailedException: line 1:1: Table 'sqlserver.dbo.lowercase' already exists
	at io.trino.testing.AbstractTestingTrinoClient.execute(AbstractTestingTrinoClient.java:138)
	at io.trino.testing.DistributedQueryRunner.executeInternal(DistributedQueryRunner.java:631)
	at io.trino.testing.DistributedQueryRunner.executeWithPlan(DistributedQueryRunner.java:613)
	at io.trino.testing.QueryAssertions.assertDistributedUpdate(QueryAssertions.java:112)
	at io.trino.testing.QueryAssertions.assertUpdate(QueryAssertions.java:66)
	at io.trino.testing.QueryAssertions.assertUpdate(QueryAssertions.java:60)
	at io.trino.testing.AbstractTestQueryFramework.assertUpdate(AbstractTestQueryFramework.java:419)
	at io.trino.testing.AbstractTestQueryFramework.assertUpdate(AbstractTestQueryFramework.java:414)
	at io.trino.plugin.sqlserver.TestSqlServerConnectorTest.testCreateAndDropTableWithSpecialCharacterName(TestSqlServerConnectorTest.java:184)

Release notes

(x) This is not user-visible or is docs only, and no release notes are required.
( ) Release notes are required. Please propose a release note for me.
( ) Release notes are required, with the following suggested text:

[chenjian2664]

we seems not run the test concurrently? It's better to share the logs of the failure test

[chenjian2664]

It's hard to read the test about the table name, does the fix of testTableNameTestData works ?

[ebyhr]

Another approach is to use AbstractTestQueryFramework#executeExclusively method. It still has a shared problem - a failure may affect other tests since the relevant test don't have finally block to drop tables.

Or just changing the prefix in each test likes #28508 might be sufficient.

[marcinsbd]

I believe those randomized suffixes are good enough.
True, we can move drop to finally, and iirc synapse does not support "IF EXIST" in drop statement

[marcinsbd]

@LBPeraza is adding randomized suffix connected with running concurrently profiles|suites: weekly and nightly?

[ebyhr]

@chenjian2664 It runs concurrently since AbstractTestQueryFramework has @TestInstance(PER_CLASS) & @Execution(CONCURRENT).

[chenjian2664]

@ebyhr Thanks for the explanation

[ebyhr]

There are 3 usages of testTableNameTestData, and each test has a different prefix:

• testCreateAndDropTableWithSpecialCharacterName: (no prefix)
• testRenameColumnNameAdditionalTests: tcn_
• testRenameFromToTableWithSpecialCharacterName: test_rename_source_

I'm not sure where the conflict is occurring. TestSqlServerConnectorTest vs BaseConnectorTest?

[LBPeraza]

we seems not run the test concurrently? It's better to share the logs of the failure test

I've added a failure stack trace to the PR description.

There 3 usages of testTableNameTestData, and each test has a different prefix:

• testCreateAndDropTableWithSpecialCharacterName: (no prefix)
• testRenameColumnNameAdditionalTests: tcn_
• testRenameFromToTableWithSpecialCharacterName: test_rename_source_

I'm not sure where the conflict is occurring. TestSqlServerConnectorTest vs BaseConnectorTest?

testRenameFromToTableWithSpecialCharacterName initially creates the table with a test_rename_source_ prefix, but then renames it to the name without a prefix, which conflicts with the table created in testCreateAndDropTableWithSpecialCharacterName.