[fix](agg) fix be crash when double type be key in agg table

[LuGuangming]

What problem does this PR solve?

Issue Number: close #61797

Related PR: #xxx

Problem Summary:
be crash when double type be key in agg table

Release note

None

Check List (For Author)

• Test

  • Regression test
  • Unit Test
  • Manual test (add detailed scripts or steps below)
  • No need to test or manual test. Explain why:
    • This is a refactor/code format and no logic has been changed.
    • Previous test can cover this change.
    • No code files have been changed.
    • Other reason

• Behavior changed:

  • No.
  • Yes.

• Does this need documentation?

  • No.
  • Yes.

Check List (For Reviewer who merge this PR)

• Confirm the release note
• Confirm test cases
• Confirm document
• Add branch pick label

[Thearas]

Thank you for your contribution to Apache Doris.
Don't know what should be done next? See How to process your PR.

Please clearly describe your PR:

1. What problem was fixed (it's best to include specific error reporting information). How it was fixed.
2. Which behaviors were modified. What was the previous behavior, what is it now, why was it modified, and what possible impacts might there be.
3. What features were added. Why was this function added?
4. Which code was refactored and why was this part of the code refactored?
5. Which functions were optimized and what is the difference before and after the optimization?

[Vallishp]

run buildall

[yiguolei]

Is this problem only exist in 2.1? what about 3.1 4.0 4.1?

[LuGuangming]

all branch are exists, I'll backport others branch

[liutang123]

/review

[github-actions]

Code Review Summary

This PR adds logic to AddColumnsClause.analyze() to automatically mark columns as key columns in AGG_KEYS tables when no aggregate type is specified, mirroring existing logic in AddColumnClause. While the intent is correct (fixing a BE crash), the implementation has two issues that need to be addressed:

Critical Issues

1. Missing setKeysType() call (Bug): Both AddColumnClause (line 83-85) and ModifyColumnClause (line 83) call colDef.setKeysType(((OlapTable) table).getKeysType()) after the setIsKey() block. This PR omits this call. The keysType field in ColumnDef is used during analyze() (line 361) to validate complex types (BITMAP, HLL, QUANTILE_STATE). Without setting it, keysType remains null, which causes the condition keysType == null || keysType == KeysType.AGG_KEYS to always be true, potentially producing incorrect validation errors for non-AGG_KEYS tables.

2. Table lookup inside the loop (Performance): The Env.getCurrentInternalCatalog().getDbOrAnalysisException(...).getTableOrAnalysisException(...) call is placed inside the for loop, causing a redundant database+table lookup for every column in the list. This should be hoisted before the loop since tableName is invariant across iterations.

Minor Issues

3. Inconsistent exception method: Uses getDbOrAnalysisException/getTableOrAnalysisException while the sibling classes (AddColumnClause, ModifyColumnClause) use getDbOrDdlException/getTableOrDdlException.

4. Missing regression test: The PR description checks the "Regression test" box but no test file is included in the changeset.

Checklist

• Correctness: Issues found (missing setKeysType, table lookup in loop)
• Performance: Table lookup should be moved outside the loop
• Test coverage: Missing despite checkbox being ticked

[github-actions]

Bug: Missing setKeysType() call + Performance: Table lookup should be outside the loop.

Comparing with AddColumnClause.analyze() (lines 76-86), there are two problems:

1. Missing setKeysType(): AddColumnClause and ModifyColumnClause both call colDef.setKeysType(((OlapTable) table).getKeysType()) after the setIsKey() block. Without this, ColumnDef.keysType remains null, which affects the validation at ColumnDef.java:361 where keysType == null is treated the same as AGG_KEYS, potentially causing incorrect errors for complex types in non-AGG tables.

2. Table lookup in loop: The table lookup is inside the for loop but tableName doesn't change between iterations. This causes redundant DB+table lookups for each column.

Suggested fix:

if (tableName != null) {
    Table table = Env.getCurrentInternalCatalog().getDbOrAnalysisException(tableName.getDb())
            .getTableOrAnalysisException(tableName.getTbl());
    if (table instanceof OlapTable) {
        OlapTable olapTable = (OlapTable) table;
        for (ColumnDef colDef : columnDefs) {
            if (olapTable.getKeysType() == KeysType.AGG_KEYS
                    && colDef.getAggregateType() == null) {
                colDef.setIsKey(true);
            }
            colDef.setKeysType(olapTable.getKeysType());
        }
    }
}
for (ColumnDef colDef : columnDefs) {
    colDef.analyze(true);
    ...
}

Or alternatively, keep the single loop but move the table lookup before it.

[yiguolei]

skip buildall

[github-actions]

PR approved by at least one committer and no changes requested.

[github-actions]

PR approved by anyone and no changes requested.