is_exclusive in column_array/map/struct has wrong semantic , we should
make sure it's nested column is also is_exclusive which can make
behavior right in some operator like join.
## Proposed changes
backport: https://github.com/apache/doris/pull/40398
Issue Number: close #xxx
<!--Describe your changes.-->
pick from master #40776
for pass test case, also fix errors in computeResultInFe
computeResultInFe will generate wrong result set if output does not
match between final result and the node executing computeResultInFe
## Proposed changes
pick: #40271
1. Change the standard for cumu compaction capability. Tablets under
state `running` or `not ready` are capable to do cumulative compaction.
2. Abort a compaction task at the beginning when the tablet is no more
capable to do compaction.
pick: #40552
Capture rowset splits and delete predicates atomicly in
`ParallelScannerBuilder::_load` as a single read source.
In this way, we could prevent reading stale rowsets with the delete
predicates eliminated by (base) compaction.
…
## Proposed changes
```
JobExecutionConfiguration jobConfig = new JobExecutionConfiguration();
jobConfig.setExecuteType(JobExecuteType.INSTANT);
setJobConfig(jobConfig);
```
- Replaying logs should not modify the original information of the job
- Use the new optimizer to check whether the executed statement is legal
(cherry picked from commit de90051162de7004cf171bbf4d21bd95ff9f3540)
## Proposed changes
Issue Number: #40474
pick (#40671)
This PR fixes the issue where the precision of default value expressions
was incorrectly set to `0` during deserialization. The `gsonPostProcess`
method no longer assigns a default precision of `0` if the precision is
not provided. Additionally, logic has been added to ensure that
expressions without precision do not incorrectly apply a precision of
`0`.
Some dependencies in these packages might conflict with the core, as the
core already includes these two packages. Therefore, exclude
Hive-related packages from the build.
## Proposed changes
Issue Number: close #xxx
<!--Describe your changes.-->
(cherry picked from commit 70922604242b0dca279682906b70b71328805b24)
## Proposed changes
Issue Number: #40757
<!--Describe your changes.-->
## Proposed changes
Issue Number: close #xxx
<!--Describe your changes.-->
(cherry picked from commit a692af226a8292d1b0a409c914d616157b64c82a)
## Proposed changes
Issue Number: close #xxx
<!--Describe your changes.-->
pick from master #40467
- http_stream TVF should always generate one fragment plan
- http_stream TVF plan should not check root as scan node
- distinguish group_commit TVF with normal insert statement
- index and generate slot should based on type cast base slot
- agg_state could cast from nullable to non-nullable
- colocated and bucket scan range compute should only on scan node
