[Fix](topn opt) prevent from merge __TEMP__ column in segment iterator

[eldenmoon]

Proposed changes

Issue Number: close #xxx

Further comments

If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...

[eldenmoon]

run buildall

[eldenmoon]

it's hard to reproduce and write regression-test， so i missed it

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[doris-robot]

TeamCity be ut coverage result:
Function Coverage: 36.61% (8585/23448)
Line Coverage: 28.68% (69803/243349)
Region Coverage: 27.65% (36073/130454)
Branch Coverage: 24.36% (18424/75618)
Coverage Report: http://coverage.selectdb-in.cc/coverage/387fed62c870dbacf2a8bc0c06ce7971f32988b9_387fed62c870dbacf2a8bc0c06ce7971f32988b9/report/index.html

[doris-robot]

TPC-H test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G'

Tpch sf100 test result on commit 387fed62c870dbacf2a8bc0c06ce7971f32988b9, data reload: false

run tpch-sf100 query with default conf and session variables
q1	4753	4501	4495	4495
q2	365	181	158	158
q3	1489	1358	1269	1269
q4	1144	976	906	906
q5	3164	3163	3168	3163
q6	259	136	137	136
q7	1037	510	490	490
q8	2250	2283	2238	2238
q9	6740	6683	6719	6683
q10	3226	3299	3276	3276
q11	336	208	213	208
q12	351	202	207	202
q13	4569	3808	3773	3773
q14	242	213	212	212
q15	575	525	522	522
q16	448	377	380	377
q17	1033	724	552	552
q18	7117	6853	6803	6803
q19	1626	1638	1641	1638
q20	558	295	318	295
q21	3184	2705	2799	2705
q22	367	302	310	302
Total cold run time: 44833 ms
Total hot run time: 40403 ms

run tpch-sf100 query with default conf and set session variable runtime_filter_mode=off
q1	4390	4422	4409	4409
q2	272	168	167	167
q3	3515	3522	3513	3513
q4	2441	2427	2426	2426
q5	5731	5735	5742	5735
q6	248	126	126	126
q7	2390	1860	1838	1838
q8	3601	3620	3605	3605
q9	9012	8959	8954	8954
q10	3932	4010	4029	4010
q11	495	378	385	378
q12	776	603	603	603
q13	4270	3558	3563	3558
q14	286	255	253	253
q15	580	516	519	516
q16	492	466	461	461
q17	1972	1958	1952	1952
q18	8751	8298	8293	8293
q19	1838	1830	1834	1830
q20	2276	1958	1962	1958
q21	6617	6252	6211	6211
q22	547	464	458	458
Total cold run time: 64432 ms
Total hot run time: 61254 ms

[doris-robot]

(From new machine)TeamCity pipeline, clickbench performance test result:
the sum of best hot time: 44.18 seconds
stream load tsv: 563 seconds loaded 74807831229 Bytes, about 126 MB/s
stream load json: 19 seconds loaded 2358488459 Bytes, about 118 MB/s
stream load orc: 66 seconds loaded 1101869774 Bytes, about 15 MB/s
stream load parquet: 32 seconds loaded 861443392 Bytes, about 25 MB/s
insert into select: 28.5 seconds inserted 10000000 Rows, about 350K ops/s
storage size: 17183730718 Bytes

[dataroaring]

LGTM

[github-actions]

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

[github-actions]

PR approved by anyone and no changes requested.

[xiaokang]

LGTM