…so, (apache#27469)

1. if skip compile then skip p0 p1 external pipelinex_p0 also
2. if change files too many, export all_files may cause problem of 'Argument list too long', change to use file
Co-authored-by: stephen <hello-stephen@qq.com>

… sort slot (apache#27137)

…window function (apache#26899)

… in agg node (apache#27282)

…ase table (apache#27387)

normally, mv column's data type should be same as base table. This pr plays as a fail-safe, if mv column's data type is different from base table accidentally, fall back to select base table to make the query works.

use batching to optimize auto partition

1. Fix a profile bug of `MergeRangeFileReader`, and add a profile `ApplyBytes` to show the total bytes  of ranges.
2. There's no need to merge large columns, because `MergeRangeFileReader` will increase the copy time.

… line (apache#27340)

…7327)

For hive version lower than 2.3.7, there is no enum ClientCapability.INSERT_ONLY_TABLES.
So if we send this enum to the server side, the server side will get a null,
and this will cause some undefined behavior, eg, failed to get tables infos from hms.

…oid concurrent delete (apache#27484)

…27475)

…ache#27467)

Finished job must be removed from the job list, otherwise the next batch of jobs will not be scheduled.

…by default (apache#27482)

Add `Eager Aggregate` rules into Rewrite rules.

…#27411)