You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@doris.apache.org by "airborne12 (via GitHub)" <gi...@apache.org> on 2023/11/09 10:53:55 UTC

[PR] [Improvement](segment iterator) remove range in first read to save time [doris]

airborne12 opened a new pull request, #26689:
URL: https://github.com/apache/doris/pull/26689

   ## Proposed changes
   
   Issue Number: close #xxx
   
   <!--Describe your changes.-->
   
   ## Further comments
   
   If this is a relatively large or complex change, kick off the discussion at [dev@doris.apache.org](mailto:dev@doris.apache.org) by explaining why you chose the solution you did and what alternatives you considered, etc...
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "qidaye (via GitHub)" <gi...@apache.org>.
qidaye merged PR #26689:
URL: https://github.com/apache/doris/pull/26689


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on code in PR #26689:
URL: https://github.com/apache/doris/pull/26689#discussion_r1390227681


##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -189,9 +193,38 @@ class SegmentIterator::BackwardBitmapRangeIterator : public SegmentIterator::Bit
 
         return true;
     }
+    uint32_t read_batch_rowids(rowid_t* buf, uint32_t batch_size) override {
+        if (!_riter.has_value || _rowid_left == 0) {
+            return 0;
+        }
+
+        if (_rowid_count <= batch_size) {
+            roaring_bitmap_to_uint32_array(_riter.parent,
+                                           buf); // Fill 'buf' with '_rowid_count' elements.
+            uint32_t num_read = _rowid_left;     // Save the number of row IDs read.
+            _rowid_left = 0;                     // No row IDs left after this operation.
+            return num_read;                     // Return the number of row IDs read.
+        }
+
+        uint32_t read_size = std::min(batch_size, _rowid_left);
+        uint32_t num_read = 0; // Counter for the number of row IDs read.
+
+        // Read row IDs into the buffer in reverse order.
+        while (num_read < read_size && _riter.has_value) {
+            buf[read_size - num_read - 1] = _riter.current_value;
+            num_read++;
+            _rowid_left--; // Decrement the count of remaining row IDs.
+            roaring_previous_uint32_iterator(&_riter);

Review Comment:
   could not found batch function



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1805059263

   (From new machine)TeamCity pipeline, clickbench performance test result:
    the sum of best hot time: 44.92 seconds
    stream load tsv:          558 seconds loaded 74807831229 Bytes, about 127 MB/s
    stream load json:         20 seconds loaded 2358488459 Bytes, about 112 MB/s
    stream load orc:          65 seconds loaded 1101869774 Bytes, about 16 MB/s
    stream load parquet:          32 seconds loaded 861443392 Bytes, about 25 MB/s
    insert into select:          28.9 seconds inserted 10000000 Rows, about 346K ops/s
    storage size: 17162276631 Bytes


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804967208

   clang-tidy review says "All clean, LGTM! :+1:"


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806851043

   TeamCity be ut coverage result:
    Function Coverage: 36.78% (8408/22863) 
    Line Coverage: 29.28% (68142/232757)
    Region Coverage: 27.90% (35228/126260)
    Branch Coverage: 24.71% (18010/72872)
    Coverage Report: http://coverage.selectdb-in.cc/coverage/e09def9222dd6308179f0007d84fcf41e6c15c9a_e09def9222dd6308179f0007d84fcf41e6c15c9a/report/index.html


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804123760

   
   <details>
   <summary>TPC-H test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G'</summary>
   
   ```
   Tpch sf100 test result on commit 9b13941ca68c11f0c290ab745ad05713315163b9, data reload: false
   
   run tpch-sf100 query with default conf and session variables
   q1	5273	5132	5282	5132
   q2	389	205	198	198
   q3	2091	2094	2069	2069
   q4	1477	1435	1439	1435
   q5	4092	4170	4063	4063
   q6	251	133	134	133
   q7	2059	1602	1621	1602
   q8	2756	2759	2741	2741
   q9	10510	10411	10216	10216
   q10	3485	3568	3562	3562
   q11	374	234	258	234
   q12	446	308	293	293
   q13	4537	4133	4102	4102
   q14	332	297	301	297
   q15	610	554	595	554
   q16	693	624	595	595
   q17	1144	1098	1095	1095
   q18	7950	7492	7528	7492
   q19	1736	1686	1700	1686
   q20	595	357	356	356
   q21	4905	4600	4589	4589
   q22	517	422	433	422
   Total cold run time: 56222 ms
   Total hot run time: 52866 ms
   
   run tpch-sf100 query with default conf and set session variable runtime_filter_mode=off
   q1	5038	5077	5248	5077
   q2	370	264	248	248
   q3	4075	4087	3904	3904
   q4	2801	2763	2764	2763
   q5	6585	6482	6584	6482
   q6	244	132	129	129
   q7	3171	2724	2724	2724
   q8	4895	4853	4865	4853
   q9	17979	17832	17847	17832
   q10	4090	4140	4157	4140
   q11	770	656	656	656
   q12	995	833	828	828
   q13	4324	3881	3904	3881
   q14	385	351	355	351
   q15	612	589	564	564
   q16	780	685	709	685
   q17	3985	3929	3858	3858
   q18	9610	9352	9319	9319
   q19	1867	1782	1808	1782
   q20	2419	2085	2097	2085
   q21	8858	8822	8753	8753
   q22	951	872	855	855
   Total cold run time: 84804 ms
   Total hot run time: 81769 ms
   ```
   </details>
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806851980

   
   <details>
   <summary>TPC-H test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G'</summary>
   
   ```
   Tpch sf100 test result on commit e09def9222dd6308179f0007d84fcf41e6c15c9a, data reload: false
   
   run tpch-sf100 query with default conf and session variables
   q1	5257	5166	5146	5146
   q2	369	185	200	185
   q3	2085	2040	2025	2025
   q4	1460	1444	1434	1434
   q5	4096	4184	4117	4117
   q6	255	134	129	129
   q7	2071	1600	1603	1600
   q8	2758	2739	2750	2739
   q9	10420	10393	10268	10268
   q10	3505	3559	3559	3559
   q11	364	245	244	244
   q12	457	297	301	297
   q13	4527	4121	4132	4121
   q14	321	288	290	288
   q15	654	572	589	572
   q16	697	615	595	595
   q17	1151	1117	1095	1095
   q18	7966	7500	7447	7447
   q19	1683	1707	1665	1665
   q20	603	362	366	362
   q21	4926	4585	4555	4555
   q22	534	417	424	417
   Total cold run time: 56159 ms
   Total hot run time: 52860 ms
   
   run tpch-sf100 query with default conf and set session variable runtime_filter_mode=off
   q1	5079	5040	5060	5040
   q2	352	235	242	235
   q3	4006	4016	3954	3954
   q4	2762	2736	2735	2735
   q5	6654	6614	6617	6614
   q6	242	124	129	124
   q7	3235	2744	2756	2744
   q8	4877	4849	4864	4849
   q9	17901	17739	17819	17739
   q10	4120	4175	4168	4168
   q11	737	666	651	651
   q12	1006	826	823	823
   q13	4297	3889	3887	3887
   q14	372	344	365	344
   q15	622	567	571	567
   q16	800	679	715	679
   q17	3925	3875	3928	3875
   q18	9582	9320	9138	9138
   q19	1841	1771	1785	1771
   q20	2434	2096	2046	2046
   q21	8949	8704	8700	8700
   q22	909	884	854	854
   Total cold run time: 84702 ms
   Total hot run time: 81537 ms
   ```
   </details>
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806846701

   run buildall


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804962817

   run buildall


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1803613911

   clang-tidy review says "All clean, LGTM! :+1:"


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804089716

   TeamCity be ut coverage result:
    Function Coverage: 36.75% (8396/22849) 
    Line Coverage: 29.24% (68071/232773)
    Region Coverage: 27.88% (35209/126268)
    Branch Coverage: 24.69% (18002/72916)
    Coverage Report: http://coverage.selectdb-in.cc/coverage/9b13941ca68c11f0c290ab745ad05713315163b9_9b13941ca68c11f0c290ab745ad05713315163b9/report/index.html


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806849975

   > could we provide any performance result?
   
   updated


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1805067646

   
   <details>
   <summary>TPC-H test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G'</summary>
   
   ```
   Tpch sf100 test result on commit 868c6b6e46dfaf73dd7644b8b1c4b51e5b58f5bf, data reload: false
   
   run tpch-sf100 query with default conf and session variables
   q1	5246	5253	5076	5076
   q2	388	209	197	197
   q3	2090	2062	2046	2046
   q4	1474	1410	1442	1410
   q5	4140	4151	4092	4092
   q6	251	134	129	129
   q7	2074	1594	1598	1594
   q8	2789	2764	2783	2764
   q9	10664	10416	10357	10357
   q10	3498	3575	3567	3567
   q11	371	252	245	245
   q12	451	301	289	289
   q13	4529	4100	4125	4100
   q14	328	293	293	293
   q15	659	568	561	561
   q16	709	625	595	595
   q17	1152	1115	1085	1085
   q18	7840	7585	7521	7521
   q19	1709	1726	1726	1726
   q20	597	359	357	357
   q21	4981	4627	4600	4600
   q22	545	451	434	434
   Total cold run time: 56485 ms
   Total hot run time: 53038 ms
   
   run tpch-sf100 query with default conf and set session variable runtime_filter_mode=off
   q1	5038	5007	5112	5007
   q2	324	237	233	233
   q3	4088	4000	3970	3970
   q4	2786	2767	2733	2733
   q5	6692	6697	6634	6634
   q6	257	127	127	127
   q7	3228	2758	2744	2744
   q8	4910	4893	4866	4866
   q9	17856	17620	17767	17620
   q10	4061	4173	4173	4173
   q11	764	655	643	643
   q12	989	816	825	816
   q13	4296	3876	3879	3876
   q14	401	375	372	372
   q15	645	576	581	576
   q16	780	706	681	681
   q17	4025	3992	3942	3942
   q18	9501	9372	9348	9348
   q19	1867	1770	1806	1770
   q20	2395	2081	2037	2037
   q21	8913	8686	8675	8675
   q22	955	872	878	872
   Total cold run time: 84771 ms
   Total hot run time: 81715 ms
   ```
   </details>
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806844081

   run buildall


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1803605097

   run buildall


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804055553

   run buildall


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "zzzxl1993 (via GitHub)" <gi...@apache.org>.
zzzxl1993 commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806696548

   LGTM


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on code in PR #26689:
URL: https://github.com/apache/doris/pull/26689#discussion_r1390243461


##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -2127,35 +2205,28 @@
     }
 }
 
-void SegmentIterator::_build_index_result_column(uint16_t* sel_rowid_idx, uint16_t select_size,
-                                                 vectorized::Block* block,
+void SegmentIterator::_build_index_result_column(const uint16_t* sel_rowid_idx,

Review Comment:
   warning: method '_build_index_result_column' can be made const [readability-make-member-function-const]
   
   be/src/olap/rowset/segment_v2/segment_iterator.cpp:2210:
   ```diff
   -                                                  const roaring::Roaring& index_result) {
   +                                                  const roaring::Roaring& index_result) const {
   ```
   



##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -99,11 +99,12 @@ class SegmentIterator::BitmapRangeIterator {
 
     explicit BitmapRangeIterator(const roaring::Roaring& bitmap) {
         roaring_init_iterator(&bitmap.roaring, &_iter);
-        _read_next_batch();
     }
 
     bool has_more_range() const { return !_eof; }
 
+    [[nodiscard]] uint32_t get_batch_size() const { return kBatchSize; }

Review Comment:
   warning: method 'get_batch_size' can be made static [readability-convert-member-functions-to-static]
   
   ```suggestion
       [[nodiscard]] static uint32_t get_batch_size() { return kBatchSize; }
   ```
   



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804058780

   clang-tidy review says "All clean, LGTM! :+1:"


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1805016993

   TeamCity be ut coverage result:
    Function Coverage: 36.76% (8403/22858) 
    Line Coverage: 29.26% (68101/232731)
    Region Coverage: 27.89% (35209/126223)
    Branch Coverage: 24.71% (18009/72880)
    Coverage Report: http://coverage.selectdb-in.cc/coverage/868c6b6e46dfaf73dd7644b8b1c4b51e5b58f5bf_868c6b6e46dfaf73dd7644b8b1c4b51e5b58f5bf/report/index.html


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806652984

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


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "airborne12 (via GitHub)" <gi...@apache.org>.
airborne12 commented on code in PR #26689:
URL: https://github.com/apache/doris/pull/26689#discussion_r1390221566


##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -189,9 +193,38 @@ class SegmentIterator::BackwardBitmapRangeIterator : public SegmentIterator::Bit
 
         return true;
     }
+    uint32_t read_batch_rowids(rowid_t* buf, uint32_t batch_size) override {
+        if (!_riter.has_value || _rowid_left == 0) {
+            return 0;
+        }
+
+        if (_rowid_count <= batch_size) {

Review Comment:
   No, here we are making a one-time determination of whether the size of the roaring bitmap is smaller than the batch_size, rather than judging the remaining size through '_rowid_left'.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806854544

   (From new machine)TeamCity pipeline, clickbench performance test result:
    the sum of best hot time: 44.44 seconds
    stream load tsv:          556 seconds loaded 74807831229 Bytes, about 128 MB/s
    stream load json:         20 seconds loaded 2358488459 Bytes, about 112 MB/s
    stream load orc:          66 seconds loaded 1101869774 Bytes, about 15 MB/s
    stream load parquet:          34 seconds loaded 861443392 Bytes, about 24 MB/s
    insert into select:          28.7 seconds inserted 10000000 Rows, about 348K ops/s
    storage size: 17162186180 Bytes


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1807019289

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


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1803670587

   TeamCity be ut coverage result:
    Function Coverage: 36.73% (8392/22847) 
    Line Coverage: 29.22% (68017/232751)
    Region Coverage: 27.87% (35183/126244)
    Branch Coverage: 24.68% (17989/72902)
    Coverage Report: http://coverage.selectdb-in.cc/coverage/8c74f262b939a6f80d1b0fdec63f11d454cd77ef_8c74f262b939a6f80d1b0fdec63f11d454cd77ef/report/index.html


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1803712974

   (From new machine)TeamCity pipeline, clickbench performance test result:
    the sum of best hot time: 44.93 seconds
    stream load tsv:          555 seconds loaded 74807831229 Bytes, about 128 MB/s
    stream load json:         20 seconds loaded 2358488459 Bytes, about 112 MB/s
    stream load orc:          65 seconds loaded 1101869774 Bytes, about 16 MB/s
    stream load parquet:          32 seconds loaded 861443392 Bytes, about 25 MB/s
    insert into select:          28.9 seconds inserted 10000000 Rows, about 346K ops/s
    storage size: 17162102082 Bytes


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1803756382

   
   <details>
   <summary>TPC-H test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G'</summary>
   
   ```
   Tpch sf100 test result on commit 8c74f262b939a6f80d1b0fdec63f11d454cd77ef, data reload: false
   
   run tpch-sf100 query with default conf and session variables
   q1	5355	5100	5197	5100
   q2	378	187	197	187
   q3	2074	2111	2037	2037
   q4	1473	1440	1443	1440
   q5	4142	4203	4107	4107
   q6	251	134	132	132
   q7	2097	1629	1621	1621
   q8	2773	2745	2746	2745
   q9	10849	10316	10199	10199
   q10	3517	3575	3572	3572
   q11	371	263	256	256
   q12	458	296	299	296
   q13	4548	4153	4105	4105
   q14	327	292	295	292
   q15	636	585	600	585
   q16	697	617	598	598
   q17	1142	1063	1096	1063
   q18	7880	7473	7416	7416
   q19	1701	1711	1720	1711
   q20	599	376	361	361
   q21	4932	4579	4560	4560
   q22	529	441	434	434
   Total cold run time: 56729 ms
   Total hot run time: 52817 ms
   
   run tpch-sf100 query with default conf and set session variable runtime_filter_mode=off
   q1	5078	5050	5246	5050
   q2	344	246	248	246
   q3	4080	3996	4067	3996
   q4	2791	2775	2753	2753
   q5	6622	6580	6544	6544
   q6	246	127	129	127
   q7	3205	2722	2762	2722
   q8	4857	4841	4873	4841
   q9	17887	17956	17858	17858
   q10	4099	4144	4184	4144
   q11	767	675	656	656
   q12	1017	866	871	866
   q13	4324	3942	3930	3930
   q14	386	368	352	352
   q15	618	581	583	581
   q16	757	687	714	687
   q17	3912	3871	3833	3833
   q18	9446	9302	9336	9302
   q19	1839	1789	1793	1789
   q20	2427	2074	2046	2046
   q21	8971	8739	8782	8739
   q22	943	837	859	837
   Total cold run time: 84616 ms
   Total hot run time: 81899 ms
   ```
   </details>
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "doris-robot (via GitHub)" <gi...@apache.org>.
doris-robot commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804126017

   (From new machine)TeamCity pipeline, clickbench performance test result:
    the sum of best hot time: 45.68 seconds
    stream load tsv:          558 seconds loaded 74807831229 Bytes, about 127 MB/s
    stream load json:         20 seconds loaded 2358488459 Bytes, about 112 MB/s
    stream load orc:          65 seconds loaded 1101869774 Bytes, about 16 MB/s
    stream load parquet:          32 seconds loaded 861443392 Bytes, about 25 MB/s
    insert into select:          28.9 seconds inserted 10000000 Rows, about 346K ops/s
    storage size: 17162208064 Bytes


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "xiaokang (via GitHub)" <gi...@apache.org>.
xiaokang commented on code in PR #26689:
URL: https://github.com/apache/doris/pull/26689#discussion_r1390101855


##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -1614,53 +1647,27 @@ Status SegmentIterator::_read_columns_by_index(uint32_t nrows_read_limit, uint32
                                                bool set_block_rowid) {
     SCOPED_RAW_TIMER(&_opts.stats->first_read_ns);
 
-    do {
-        uint32_t range_from = 0;
-        uint32_t range_to = 0;
-        bool has_next_range =
-                _range_iter->next_range(nrows_read_limit - nrows_read, &range_from, &range_to);
-        if (!has_next_range) {
-            break;
-        }
-
-        size_t rows_to_read = range_to - range_from;
-        _cur_rowid = range_to;
-
-        if (set_block_rowid) {
-            // Here use std::iota is better performance than for-loop, maybe for-loop is not vectorized
-            auto start = _block_rowids.data() + nrows_read;
-            auto end = start + rows_to_read;
-            std::iota(start, end, range_from);
-            nrows_read += rows_to_read;
-        } else {
-            nrows_read += rows_to_read;
-        }
-
-        _split_row_ranges.emplace_back(std::pair {range_from, range_to});
-    } while (nrows_read < nrows_read_limit && !_opts.read_orderby_key_reverse);
+    nrows_read = _range_iter->read_batch_rowids(_block_rowids.data(), nrows_read_limit);
+    bool is_continuous = (nrows_read > 1) &&
+                         (_block_rowids[nrows_read - 1] - _block_rowids[0] == nrows_read - 1);
 
     for (auto cid : _first_read_column_ids) {
         auto& column = _current_return_columns[cid];
         if (_prune_column(cid, column, true, nrows_read)) {
             continue;
         }
-        for (auto& range : _split_row_ranges) {
-            size_t nrows = range.second - range.first;
-            {
-                _opts.stats->block_first_read_seek_num += 1;
-                if (_opts.runtime_state && _opts.runtime_state->enable_profile()) {
-                    SCOPED_RAW_TIMER(&_opts.stats->block_first_read_seek_ns);
-                    RETURN_IF_ERROR(_column_iterators[cid]->seek_to_ordinal(range.first));
-                } else {
-                    RETURN_IF_ERROR(_column_iterators[cid]->seek_to_ordinal(range.first));
-                }
-            }
-            size_t rows_read = nrows;
+
+        if (is_continuous) {
+            size_t rows_read = nrows_read;
+            RETURN_IF_ERROR(_column_iterators[cid]->seek_to_ordinal(_block_rowids[0]));

Review Comment:
   add profile as before



##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -141,6 +140,9 @@ class SegmentIterator::BitmapRangeIterator {
         *to = *from + range_size;
         return true;
     }
+    virtual uint32_t read_batch_rowids(rowid_t* buf, uint32_t batch_size) {

Review Comment:
   add comment for the sematics



##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -141,6 +140,9 @@ class SegmentIterator::BitmapRangeIterator {
         *to = *from + range_size;
         return true;
     }
+    virtual uint32_t read_batch_rowids(rowid_t* buf, uint32_t batch_size) {

Review Comment:
   btw, add a blank line



##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -2138,22 +2143,21 @@ void SegmentIterator::_build_index_result_column(uint16_t* sel_rowid_idx, uint16
     size_t idx_in_row_range = 0;
     size_t idx_in_selected = 0;
     // _split_row_ranges store multiple ranges which split in function _read_columns_by_index(),
-    // index_result is a column predicate apply result in a whole segement,
-    // but a scanner thread one time can read max rows limit by block_row_max,
+    // index_result is a column predicate apply result in a whole segment,
+    // but a scanner thread can read max rows limit by block_row_max one time,
     // so split _row_bitmap by one time scan range, in order to match size of one scanner thread read rows.
-    for (auto origin_row_range : _split_row_ranges) {
-        for (size_t rowid = origin_row_range.first; rowid < origin_row_range.second; ++rowid) {
-            if (sel_rowid_idx == nullptr || (idx_in_selected < select_size &&
-                                             idx_in_row_range == sel_rowid_idx[idx_in_selected])) {
-                if (index_result.contains(rowid)) {
-                    vec_match_pred[idx_in_block++] = true;
-                } else {
-                    vec_match_pred[idx_in_block++] = false;
-                }
-                idx_in_selected++;
+    for (uint32_t idx = 0; idx < _current_batch_rows_read; idx++) {
+        auto rowid = _block_rowids[idx];
+        if (sel_rowid_idx == nullptr ||
+            (idx_in_selected < select_size && idx_in_row_range == sel_rowid_idx[idx_in_selected])) {

Review Comment:
   The original logic `idx_in_row_range == sel_rowid_idx[idx_in_selected])` indicate that data in sel_rowid_idx is relative index in read block. It need to be verified once more.



##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -1614,53 +1647,27 @@ Status SegmentIterator::_read_columns_by_index(uint32_t nrows_read_limit, uint32
                                                bool set_block_rowid) {
     SCOPED_RAW_TIMER(&_opts.stats->first_read_ns);
 
-    do {
-        uint32_t range_from = 0;
-        uint32_t range_to = 0;
-        bool has_next_range =
-                _range_iter->next_range(nrows_read_limit - nrows_read, &range_from, &range_to);
-        if (!has_next_range) {
-            break;
-        }
-
-        size_t rows_to_read = range_to - range_from;
-        _cur_rowid = range_to;
-
-        if (set_block_rowid) {
-            // Here use std::iota is better performance than for-loop, maybe for-loop is not vectorized
-            auto start = _block_rowids.data() + nrows_read;
-            auto end = start + rows_to_read;
-            std::iota(start, end, range_from);
-            nrows_read += rows_to_read;
-        } else {
-            nrows_read += rows_to_read;
-        }
-
-        _split_row_ranges.emplace_back(std::pair {range_from, range_to});
-    } while (nrows_read < nrows_read_limit && !_opts.read_orderby_key_reverse);
+    nrows_read = _range_iter->read_batch_rowids(_block_rowids.data(), nrows_read_limit);
+    bool is_continuous = (nrows_read > 1) &&
+                         (_block_rowids[nrows_read - 1] - _block_rowids[0] == nrows_read - 1);
 
     for (auto cid : _first_read_column_ids) {
         auto& column = _current_return_columns[cid];
         if (_prune_column(cid, column, true, nrows_read)) {
             continue;
         }
-        for (auto& range : _split_row_ranges) {
-            size_t nrows = range.second - range.first;
-            {
-                _opts.stats->block_first_read_seek_num += 1;
-                if (_opts.runtime_state && _opts.runtime_state->enable_profile()) {
-                    SCOPED_RAW_TIMER(&_opts.stats->block_first_read_seek_ns);
-                    RETURN_IF_ERROR(_column_iterators[cid]->seek_to_ordinal(range.first));
-                } else {
-                    RETURN_IF_ERROR(_column_iterators[cid]->seek_to_ordinal(range.first));
-                }
-            }
-            size_t rows_read = nrows;
+
+        if (is_continuous) {
+            size_t rows_read = nrows_read;
+            RETURN_IF_ERROR(_column_iterators[cid]->seek_to_ordinal(_block_rowids[0]));
             RETURN_IF_ERROR(_column_iterators[cid]->next_batch(&rows_read, column));
-            if (rows_read != nrows) {
-                return Status::Error<ErrorCode::INTERNAL_ERROR>("nrows({}) != rows_read({})", nrows,
-                                                                rows_read);
+            if (rows_read != nrows_read) {
+                return Status::Error<ErrorCode::INTERNAL_ERROR>("nrows({}) != rows_read({})",
+                                                                nrows_read, rows_read);
             }
+        } else {
+            RETURN_IF_ERROR(_column_iterators[cid]->read_by_rowids(_block_rowids.data(), nrows_read,

Review Comment:
   add profile



##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -2138,22 +2143,21 @@ void SegmentIterator::_build_index_result_column(uint16_t* sel_rowid_idx, uint16
     size_t idx_in_row_range = 0;
     size_t idx_in_selected = 0;
     // _split_row_ranges store multiple ranges which split in function _read_columns_by_index(),
-    // index_result is a column predicate apply result in a whole segement,
-    // but a scanner thread one time can read max rows limit by block_row_max,
+    // index_result is a column predicate apply result in a whole segment,
+    // but a scanner thread can read max rows limit by block_row_max one time,
     // so split _row_bitmap by one time scan range, in order to match size of one scanner thread read rows.
-    for (auto origin_row_range : _split_row_ranges) {
-        for (size_t rowid = origin_row_range.first; rowid < origin_row_range.second; ++rowid) {
-            if (sel_rowid_idx == nullptr || (idx_in_selected < select_size &&
-                                             idx_in_row_range == sel_rowid_idx[idx_in_selected])) {
-                if (index_result.contains(rowid)) {
-                    vec_match_pred[idx_in_block++] = true;
-                } else {
-                    vec_match_pred[idx_in_block++] = false;
-                }
-                idx_in_selected++;
+    for (uint32_t idx = 0; idx < _current_batch_rows_read; idx++) {

Review Comment:
   The original logic is complicated and there are too many idx. We can replace idx by i to reduce idx variables.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "xiaokang (via GitHub)" <gi...@apache.org>.
xiaokang commented on code in PR #26689:
URL: https://github.com/apache/doris/pull/26689#discussion_r1390106600


##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -189,9 +193,38 @@ class SegmentIterator::BackwardBitmapRangeIterator : public SegmentIterator::Bit
 
         return true;
     }
+    uint32_t read_batch_rowids(rowid_t* buf, uint32_t batch_size) override {
+        if (!_riter.has_value || _rowid_left == 0) {
+            return 0;
+        }
+
+        if (_rowid_count <= batch_size) {

Review Comment:
   Is it OK if _rowid_left <= batch_size?



##########
be/src/olap/rowset/segment_v2/segment_iterator.cpp:
##########
@@ -189,9 +193,38 @@ class SegmentIterator::BackwardBitmapRangeIterator : public SegmentIterator::Bit
 
         return true;
     }
+    uint32_t read_batch_rowids(rowid_t* buf, uint32_t batch_size) override {
+        if (!_riter.has_value || _rowid_left == 0) {
+            return 0;
+        }
+
+        if (_rowid_count <= batch_size) {
+            roaring_bitmap_to_uint32_array(_riter.parent,
+                                           buf); // Fill 'buf' with '_rowid_count' elements.
+            uint32_t num_read = _rowid_left;     // Save the number of row IDs read.
+            _rowid_left = 0;                     // No row IDs left after this operation.
+            return num_read;                     // Return the number of row IDs read.
+        }
+
+        uint32_t read_size = std::min(batch_size, _rowid_left);
+        uint32_t num_read = 0; // Counter for the number of row IDs read.
+
+        // Read row IDs into the buffer in reverse order.
+        while (num_read < read_size && _riter.has_value) {
+            buf[read_size - num_read - 1] = _riter.current_value;
+            num_read++;
+            _rowid_left--; // Decrement the count of remaining row IDs.
+            roaring_previous_uint32_iterator(&_riter);

Review Comment:
   Is this op time consuming in perf? May be there is more effecient function to get batch.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1806652997

   PR approved by anyone and no changes requested.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org


Re: [PR] [Improvement](segment iterator) remove range in first read to save time [doris]

Posted by "eldenmoon (via GitHub)" <gi...@apache.org>.
eldenmoon commented on PR #26689:
URL: https://github.com/apache/doris/pull/26689#issuecomment-1804983198

   could we provide any performance result?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org