PostgreSQL、KingBase 数据库 ORDER BY LIMIT 查询缓慢案例
好久没写博客了，最近从人大金仓离职了，新公司入职了蚂蚁集团，正在全力学习 OcenaBase 数据库的体系结构中。
以后分享的案例知识基本上都是以 OcenaBase 分布式数据库为主了，呦西。😁
昨天帮朋友看了个金仓KES数据库的 SQL 案例，废话不说，直接贴SQL：
慢SQL（执行时间 8s ，限制返回 30 行）
explain analyze SELECT GI.ID, GI.MODULE_ID, GI.BT, GI.WH, GI.JJCD_TEXT, GI.CREATE_DEPTNAME, GI.CREATE_TIME, GI.MODULE_NAME FROM gifgifgif GI INNER JOIN gufgufguf GUF ON (GUF.ifid = GI.ID) WHERE GI.ROWSTATE > - 1 AND (GUF.usid = '0' OR GUF.usid = '210317100256if6gVcTb3Ado1o2ytLs') AND ((GI.BT LIKE '%签%') OR (GI.MODULE_NAME LIKE '%签%') OR (GI.WH LIKE '%签%') OR (GI.JJCD_TEXT LIKE '%签%') OR (GI.CREATE_DEPTNAME LIKE '%签%')) ORDER BY GI.CREATE_TIME DESC LIMIT 30;
慢SQL执行计划
QUERY PLAN --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------- Limit (cost=1001.05..17578.06 rows=30 width=240) (actual time=6458.263..8763.733 rows=7 loops=1) -> Gather Merge (cost=1001.05..3879467.79 rows=7019 width=240) (actual time=6458.261..8763.728 rows=7 loops=1) Workers Planned: 4 Workers Launched: 4 -> Nested Loop (cost=0.99..3877631.71 rows=1755 width=240) (actual time=2843.144..8274.217 rows=1 loops=5) -> Parallel Index Scan Backward using gifgifgif_CREATE_TIME1 on gifgifgif GI (cost=0.43..1158925.09 rows=433728 width=240) (actual time=0.043..2159.037 rows=350466 loops=5) Filter: ((ROWSTATE > '-1'::numeric) AND (((BT)::text ~~ '%签%'::text) OR ((MODULE_NAME)::text ~~ '%签%'::text) OR ((WH)::text ~~ '%签%'::text) OR ((JJCD_TEXT)::text ~~ '%签%'::text) OR ((CREATE_DEPTNAME)::text ~~ '% %'::text))) Rows Removed by Filter: 423271 -> Index Only Scan using idx_gufgufguf_1_2_3 on gufgufguf GUF (cost=0.56..6.26 rows=1 width=32) (actual time=0.017..0.017 rows=0 loops=1752329) -- 慢：(1752329/5) * 0.017 / 1000 = 5.95s Index Cond: (ifid = (GI.ID)::text) Filter: (((usid)::text = '0'::text) OR ((usid)::text = '210317100256if6gVcTb3Ado1o2ytLs'::text)) Rows Removed by Filter: 3 Heap Fetches: 0 Planning Time: 0.832 ms Execution Time: 8763.803 ms (15 行记录)
我看到这计划简直无语，这种SQL不能 300 ms以内出来就绝对有问题，而且这么简单的语句都能用上并行，真的服。
Index Only Scan using idx_gufgufguf_1_2_3 on gufgufguf GUF 每个并行进程执行 5.95s 这也太拉跨了。
看执行计划基本都是用 Index Scan 或者是 Index Only Scan，但是本SQL 谓词过滤条件很多 or ，其实优化器如果执行位图扫描才是最优解计划，但是CBO偏偏没执行！！！
SQL去掉 LIMIT 30限制条件:
explain analyze SELECT GI.ID, GI.MODULE_ID, GI.BT, GI.WH, GI.JJCD_TEXT, GI.CREATE_DEPTNAME, GI.CREATE_TIME, GI.MODULE_NAME FROM gifgifgif GI INNER JOIN gufgufguf GUF ON (GUF.ifid = GI.ID) WHERE GI.ROWSTATE > - 1 AND (GUF.usid = '0' OR GUF.usid = '210317100256if6gVcTb3Ado1o2ytLs') AND ((GI.BT LIKE '%签%') OR (GI.MODULE_NAME LIKE '%签%') OR (GI.WH LIKE '%签%') OR (GI.JJCD_TEXT LIKE '%签%') OR (GI.CREATE_DEPTNAME LIKE '%签%')) ORDER BY GI.CREATE_TIME DESC ;
去掉 LIMIT 30限制条件SQL执行计划：
QUERY PLAN --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------- Gather Merge (cost=98222.89..99026.61 rows=6792 width=240) (actual time=33.640..35.974 rows=7 loops=1) Workers Planned: 3 Workers Launched: 3 -> Sort (cost=97222.85..97228.51 rows=2264 width=240) (actual time=26.724..26.725 rows=2 loops=4) Sort Key: GI.CREATE_TIME DESC Sort Method: quicksort Memory: 25kB Worker 0: Sort Method: quicksort Memory: 25kB Worker 1: Sort Method: quicksort Memory: 25kB Worker 2: Sort Method: quicksort Memory: 26kB -> Nested Loop (cost=510.90..97096.70 rows=2264 width=240) (actual time=11.118..26.693 rows=2 loops=4) -> Parallel Bitmap Heap Scan on gufgufguf GUF (cost=510.35..59045.81 rows=5049 width=32) (actual time=0.480..3.498 rows=1178 loops=4) Recheck Cond: (((usid)::text = '0'::text) OR ((usid)::text = '210317100256if6gVcTb3Ado1o2ytLs'::text)) Heap Blocks: exact=1464 -> BitmapOr (cost=510.35..510.35 rows=15652 width=0) (actual time=1.567..1.568 rows=0 loops=1) -> Bitmap Index Scan on gufgufguf_usid (cost=0.00..251.26 rows=7826 width=0) (actual time=0.022..0.022 rows=0 loops=1) Index Cond: ((usid)::text = '0'::text) -> Bitmap Index Scan on gufgufguf_usid (cost=0.00..251.26 rows=7826 width=0) (actual time=1.545..1.545 rows=4713 loops=1) Index Cond: ((usid)::text = '210317100256if6gVcTb3Ado1o2ytLs'::text) -> Index Scan using gifgifgif_PKEY1 on gifgifgif GI (cost=0.56..7.54 rows=1 width=240) (actual time=0.019..0.019 rows=0 loops=4713) Index Cond: ((ID)::text = (GUF.ifid)::text) Filter: ((ROWSTATE > '-1'::numeric) AND (((BT)::text ~~ '%签%'::text) OR ((MODULE_NAME)::text ~~ '%签%'::text) OR ((WH)::text ~~ '%签%'::text) OR ((JJCD_TEXT)::text ~~ '%签%'::text) OR ((CREATE_DEPTNAME)::text ~~ '% %'::text))) Rows Removed by Filter: 1 Planning Time: 0.815 ms Execution Time: 36.060 ms (24 行记录)
可以看到去掉LIMIT 30 以后，CBO能正常使用上 Bitmap Index Scan + BitmapOr 的查询策略，SQL只需要 36ms就能跑出结果。
PG比较牛逼的地方是B+树索引能基于SQL的查询条件，自动能转换成位图索引的查询策略。
像这种情况就简单了，只需要改变下限制SQL返回条数的逻辑即可，kingbase也兼容Oracle rownum 的语法，我们可以将上面SQL等价改成 rownum 来优化。
LIMIT 改写成 rownum ：
explain analyze SELECT * FROM ( SELECT GI.ID, GI.MODULE_ID, GI.BT, GI.WH, GI.JJCD_TEXT, GI.CREATE_DEPTNAME, GI.CREATE_TIME, GI.MODULE_NAME FROM gifgifgif GI INNER JOIN gufgufguf GUF ON (GUF.ifid = GI.ID) WHERE GI.ROWSTATE > - 1 AND (GUF.usid = '0' OR GUF.usid = '210317100256if6gVcTb3Ado1o2ytLs') AND ((GI.BT LIKE '%签%') OR (GI.MODULE_NAME LIKE '%签%') OR (GI.WH LIKE '%签%') OR (GI.JJCD_TEXT LIKE '%签%') OR (GI.CREATE_DEPTNAME LIKE '%签%')) ORDER BY GI.CREATE_TIME DESC) WHERE ROWNUM <= 30;
LIMIT 改写成 rownum 执行计划：
QUERY PLAN --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------- Count (cost=98222.89..99162.45 rows=0 width=240) (actual time=31.418..33.691 rows=7 loops=1) Stop Keys: (ROWNUM <= 30) -> Gather Merge (cost=98222.89..99026.61 rows=6792 width=240) (actual time=31.415..33.686 rows=7 loops=1) Workers Planned: 3 Workers Launched: 3 -> Sort (cost=97222.85..97228.51 rows=2264 width=240) (actual time=26.497..26.498 rows=2 loops=4) Sort Key: GI.CREATE_TIME DESC Sort Method: quicksort Memory: 25kB Worker 0: Sort Method: quicksort Memory: 25kB Worker 1: Sort Method: quicksort Memory: 27kB Worker 2: Sort Method: quicksort Memory: 25kB -> Nested Loop (cost=510.90..97096.70 rows=2264 width=240) (actual time=14.246..26.465 rows=2 loops=4) -> Parallel Bitmap Heap Scan on gufgufguf GUF (cost=510.35..59045.81 rows=5049 width=32) (actual time=0.513..3.401 rows=1178 loops=4) Recheck Cond: (((usid)::text = '0'::text) OR ((usid)::text = '210317100256if6gVcTb3Ado1o2ytLs'::text)) Heap Blocks: exact=1373 -> BitmapOr (cost=510.35..510.35 rows=15652 width=0) (actual time=1.664..1.664 rows=0 loops=1) -> Bitmap Index Scan on gufgufguf_usid (cost=0.00..251.26 rows=7826 width=0) (actual time=0.024..0.024 rows=0 loops=1) Index Cond: ((usid)::text = '0'::text) -> Bitmap Index Scan on gufgufguf_usid (cost=0.00..251.26 rows=7826 width=0) (actual time=1.639..1.639 rows=4713 loops=1) Index Cond: ((usid)::text = '210317100256if6gVcTb3Ado1o2ytLs'::text) -> Index Scan using gifgifgif_PKEY1 on gifgifgif GI (cost=0.56..7.54 rows=1 width=240) (actual time=0.019..0.019 rows=0 loops=4713) Index Cond: ((ID)::text = (GUF.ifid)::text) Filter: ((ROWSTATE > '-1'::numeric) AND (((BT)::text ~~ '%签%'::text) OR ((MODULE_NAME)::text ~~ '%签%'::text) OR ((WH)::text ~~ '%签%'::text) OR ((JJCD_TEXT)::text ~~ '%签%'::text) OR ((CREATE_DEPTNAME)::text ~~ '%签%'::text))) Rows Removed by Filter: 1 Planning Time: 0.897 ms Execution Time: 33.778 ms (26 行记录)
可以看到SQL通过将LIMIT 改写成 rownum 以后，原来执行时间 8s 降低到 33ms 就能跑出结果了，本条SQL到此已经优化完毕。
再来看个 postgresql 类似案例
慢SQL如下：
-- 慢SQL ： SELECT npspppordedddd.*, npsmainproxxss.member_num, npsmainproxxss.tools, npsmainproxxss.importance, npsmainproxxss.member_type, npsmainproxxss.machine_switch, npsmainproxxss.image_url, npsmainproxxss.sys_file_id, npsmainproxxss.work_time, npsmainproxxss.opt_place, npsmainproxxss.opt_work, npsmainproxxss.opt_work_desc, npsmainproxxss.check_item, npsmainproxxss.val_min, npsmainproxxss.val_max, npsmanscheeee.DYNAMIC, npsmanscheeee.task_policy, npsmainfssdx.package_code, npsmainfssdx.package_name, npsmainfssdx.package_desc, npsmainfssdx.location_desc, npsmainfssdx.product_line, npsmainfssdx.product_line_code, npsmainfssdx.work_area, npsmainfssdx.work_area_code, npsmainfssdx.PROCEDURE, npsmainfssdx.procedure_code, npsmanscheeee.quantum, npsmanscheeee.float_time, npsmanscheeee.sort, ccssdeewqnn.team_id, cssdsewff.pos_id, cssdsewff.pos_name, npsmainfssdx.by_sort FROM npspppordedddd LEFT JOIN cssxxxaaaset ON cssxxxaaaset.ID = npspppordedddd.target_code LEFT JOIN cssxxxaaaset AS facility_location ON facility_location.ID = cssxxxaaaset.parent_id LEFT JOIN cssxxxaaaset AS fc_tp_line ON fc_tp_line.ID = facility_location.parent_id LEFT JOIN npsmanscheeee ON npsmanscheeee.ID = npspppordedddd.schedule_id LEFT JOIN npsmainproxxss ON npsmainproxxss.ID = npsmanscheeee.maintain_project_id LEFT JOIN npsmainfssdx ON npsmainfssdx.ID = npsmanscheeee.maintain_package_id LEFT JOIN cssdsewff ON cssdsewff.device_code = npsmainfssdx.target_code LEFT JOIN ccssdeewqnn ON ccssdeewqnn.team_id = cssdsewff.team_id WHERE npsmainfssdx.target_type = 'device' AND npspppordedddd.maintain_type = 'DS' AND npspppordedddd.plan_time >= '2024-02-03' AND npspppordedddd.plan_time <= '2024-03-03' AND cssxxxaaaset.level_code = 'asset' AND cssxxxaaaset.ID IN ( 'a4fe1cbb-8f3f-499e-9ae2-f4a5301b074e', '1537e802-0458-4e75-ba53-47b4c60cc9b8', '7e7aa6c8-c2dd-4373-b793-70440b940c29', 'fb5b2c90-a7cc-4734-b0e1-518f08e8972e', 'b8929b19-cfa7-4581-8867-41d091ca3ade', '8e4559af-7755-4425-a409-e3cbe97669db', 'a50e4279-5f52-4e36-b2ed-b1de6c5086e4', 'ac94281e-6549-40db-bb7b-664c07bdb84e', '80d6a2b8-9268-446a-bb0c-b787f49b262c', 'fe473ca5-d96f-49fb-959e-b70cf15919a9', '291f2271-9645-4f0f-830b-9190786d78d3', '92c9d569-fa32-4cca-b18e-8f69a0802a3c', 'f7dab2be-5b8e-4dfd-be16-af2bd2d6c6d6', '87e1890e-9a54-4f85-8340-d341f9b9a1f3', '16032ca6-3009-48eb-be54-6b32e2dce11f', '8c410ba7-852c-4510-ad61-abb6e4e62706', 'cfdaacca-c6ca-431f-ab9c-fce60df46ee1', 'c7aa7e85-0076-4b1f-a2b6-8d49d67eefc6', 'a4ea061c-ac20-4515-b069-59ce8c8a0b1b', '624826b0-1c73-4d73-a46a-84df45336b3b', '9853c273-844c-4fd7-a64a-e27b7d44f577', 'c663f5b3-6a17-4863-aa8f-4d1c8a85200c', 'e0d499cf-9b41-4da3-b501-e29cb1dcc0a3' ) AND npsmainfssdx.tenant_code IN ( '_ADMIN', '_NA_', '59052881-df03-4c49-923f-6382b8ed138b', '573bcfa0-239a-4f0b-9105-b906db4895ec', '3264de0d-7262-4dce-8178-0112c0a1603b', '0e0208f8-8763-4735-bb2c-ea5ef8bda742', 'f333dba3-7f08-4276-8bcb-a99c6ecc5120', '4b04078d-8265-400d-a27d-e40d0f1646b4', '52e44df1-be04-4bfc-bd5f-f37dbb60252c', '6088e13a-28de-4b88-908f-247889e3d068', '213fe35d-ddce-4876-a91c-f8c292713a20', 'e5fa7501-b3cb-4c59-bb78-6f8e5cdad19b', 'acda7d56-6d0f-4302-b7f4-292fb71173ba', 'ad13e67e-8bdc-45b1-96f4-f156c9c7dfb0', '8b240bd1-bc1d-4dcf-87de-4e8e3e74d004', '69869877-f9ce-4e85-a73b-edefe702bb6c', '0ad151a9-bec9-4e23-953b-9f4c2e7be9a1' ) ORDER BY npsmanscheeee.sort,npspppordedddd.opt_shift_team ASC LIMIT 15;
上面这条慢SQL 加了limit 15 以后出不了结果，如果去掉的话，10s左右就能返回查询结果。
使用 row_number 分页方案优化SQL:
-- row_number 分页 ： SELECT * FROM ( SELECT npspppordedddd.*, npsmainproxxss.member_num, npsmainproxxss.tools, npsmainproxxss.importance, npsmainproxxss.member_type, npsmainproxxss.machine_switch, npsmainproxxss.image_url, npsmainproxxss.sys_file_id, npsmainproxxss.work_time, npsmainproxxss.opt_place, npsmainproxxss.opt_work, npsmainproxxss.opt_work_desc, npsmainproxxss.check_item, npsmainproxxss.val_min, npsmainproxxss.val_max, npsmanscheeee.DYNAMIC, npsmanscheeee.task_policy, npsmainfssdx.package_code, npsmainfssdx.package_name, npsmainfssdx.package_desc, npsmainfssdx.location_desc, npsmainfssdx.product_line, npsmainfssdx.product_line_code, npsmainfssdx.work_area, npsmainfssdx.work_area_code, npsmainfssdx.PROCEDURE, npsmainfssdx.procedure_code, npsmanscheeee.quantum, npsmanscheeee.float_time, npsmanscheeee.sort, ccssdeewqnn.team_id, cssdsewff.pos_id, cssdsewff.pos_name, npsmainfssdx.by_sort, row_number() over(ORDER BY npsmanscheeee.sort,npspppordedddd.opt_shift_team ASC) rn FROM npspppordedddd LEFT JOIN cssxxxaaaset ON cssxxxaaaset.ID = npspppordedddd.target_code LEFT JOIN cssxxxaaaset AS facility_location ON facility_location.ID = cssxxxaaaset.parent_id LEFT JOIN cssxxxaaaset AS fc_tp_line ON fc_tp_line.ID = facility_location.parent_id LEFT JOIN npsmanscheeee ON npsmanscheeee.ID = npspppordedddd.schedule_id LEFT JOIN npsmainproxxss ON npsmainproxxss.ID = npsmanscheeee.maintain_project_id LEFT JOIN npsmainfssdx ON npsmainfssdx.ID = npsmanscheeee.maintain_package_id LEFT JOIN cssdsewff ON cssdsewff.device_code = npsmainfssdx.target_code LEFT JOIN ccssdeewqnn ON ccssdeewqnn.team_id = cssdsewff.team_id WHERE npsmainfssdx.target_type = 'device' AND npspppordedddd.maintain_type = 'DS' AND npspppordedddd.plan_time >= '2024-02-03' AND npspppordedddd.plan_time <= '2024-03-03' AND cssxxxaaaset.level_code = 'asset' AND cssxxxaaaset.ID IN ( 'a4fe1cbb-8f3f-499e-9ae2-f4a5301b074e', '1537e802-0458-4e75-ba53-47b4c60cc9b8', '7e7aa6c8-c2dd-4373-b793-70440b940c29', 'fb5b2c90-a7cc-4734-b0e1-518f08e8972e', 'b8929b19-cfa7-4581-8867-41d091ca3ade', '8e4559af-7755-4425-a409-e3cbe97669db', 'a50e4279-5f52-4e36-b2ed-b1de6c5086e4', 'ac94281e-6549-40db-bb7b-664c07bdb84e', '80d6a2b8-9268-446a-bb0c-b787f49b262c', 'fe473ca5-d96f-49fb-959e-b70cf15919a9', '291f2271-9645-4f0f-830b-9190786d78d3', '92c9d569-fa32-4cca-b18e-8f69a0802a3c', 'f7dab2be-5b8e-4dfd-be16-af2bd2d6c6d6', '87e1890e-9a54-4f85-8340-d341f9b9a1f3', '16032ca6-3009-48eb-be54-6b32e2dce11f', '8c410ba7-852c-4510-ad61-abb6e4e62706', 'cfdaacca-c6ca-431f-ab9c-fce60df46ee1', 'c7aa7e85-0076-4b1f-a2b6-8d49d67eefc6', 'a4ea061c-ac20-4515-b069-59ce8c8a0b1b', '624826b0-1c73-4d73-a46a-84df45336b3b', '9853c273-844c-4fd7-a64a-e27b7d44f577', 'c663f5b3-6a17-4863-aa8f-4d1c8a85200c', 'e0d499cf-9b41-4da3-b501-e29cb1dcc0a3' ) AND npsmainfssdx.tenant_code IN ( '_ADMIN', '_NA_', '59052881-df03-4c49-923f-6382b8ed138b', '573bcfa0-239a-4f0b-9105-b906db4895ec', '3264de0d-7262-4dce-8178-0112c0a1603b', '0e0208f8-8763-4735-bb2c-ea5ef8bda742', 'f333dba3-7f08-4276-8bcb-a99c6ecc5120', '4b04078d-8265-400d-a27d-e40d0f1646b4', '52e44df1-be04-4bfc-bd5f-f37dbb60252c', '6088e13a-28de-4b88-908f-247889e3d068', '213fe35d-ddce-4876-a91c-f8c292713a20', 'e5fa7501-b3cb-4c59-bb78-6f8e5cdad19b', 'acda7d56-6d0f-4302-b7f4-292fb71173ba', 'ad13e67e-8bdc-45b1-96f4-f156c9c7dfb0', '8b240bd1-bc1d-4dcf-87de-4e8e3e74d004', '69869877-f9ce-4e85-a73b-edefe702bb6c', '0ad151a9-bec9-4e23-953b-9f4c2e7be9a1' ) ) x WHERE x.rn <= 15;
使用 row_number 分页执行计划：
Subquery Scan on x (cost=73002.03..73041.39 rows=404 width=2650) (actual time=492.419..562.760 rows=15 loops=1) Filter: (x.rn <= 15) Rows Removed by Filter: 24089 -> WindowAgg (cost=73002.03..73026.25 rows=1211 width=2650) (actual time=492.417..561.576 rows=24104 loops=1) -> Sort (cost=73002.03..73005.06 rows=1211 width=2642) (actual time=492.373..511.693 rows=24104 loops=1) Sort Key: npsmanscheeee.sort, npspppordedddd.opt_shift_team Sort Method: external merge Disk: 20776kB -> Hash Left Join (cost=229.76..72940.02 rows=1211 width=2642) (actual time=3.379..432.995 rows=24104 loops=1) Hash Cond: ((cssdsewff.team_id)::text = (ccssdeewqnn.team_id)::text) -> Hash Left Join (cost=228.34..72934.70 rows=1211 width=2529) (actual time=3.352..424.280 rows=24104 loops=1) Hash Cond: ((npsmainfssdx.target_code)::text = (cssdsewff.device_code)::text) -> Nested Loop Left Join (cost=223.99..72925.80 rows=1211 width=2473) (actual time=3.252..413.474 rows=24104 loops=1) -> Hash Join (cost=223.57..67548.64 rows=1211 width=1998) (actual time=3.236..298.888 rows=24104 loops=1) Hash Cond: ((npsmanscheeee.maintain_package_id)::text = (npsmainfssdx.id)::text) -> Nested Loop (cost=134.64..67454.55 rows=1958 width=1009) (actual time=2.159..284.665 rows=24104 loops=1) -> Nested Loop (cost=134.22..58768.62 rows=1958 width=726) (actual time=2.138..163.896 rows=24104 loops=1) -> Hash Right Join (cost=133.67..194.89 rows=12 width=39) (actual time=0.655..0.839 rows=23 loops=1) Hash Cond: ((fc_tp_line.id)::text = (facility_location.parent_id)::text) -> Index Only Scan using ccs_xdevice_parent_asset_pseudo_level_idx on cssxxxaaaset fc_tp_line (cost=0.28..58.10 rows=655 width=39) (actual time=0.011..0.096 rows=655 loops=1) Heap Fetches: 0 -> Hash (cost=133.24..133.24 rows=12 width=79) (actual time=0.616..0.620 rows=23 loops=1) Buckets: 1024 Batches: 1 Memory Usage: 11kB -> Hash Left Join (cost=66.56..133.24 rows=12 width=79) (actual time=0.571..0.611 rows=23 loops=1) Hash Cond: ((cssxxxaaaset.parent_id)::text = (facility_location.id)::text) -> Index Only Scan using ccs_xdevice_parent_asset_pseudo_level_idx on cssxxxaaaset (cost=0.28..66.79 rows=12 width=79) (actual time=0.239..0.269 rows=23 loops=1) Index Cond: (level_code = 'asset'::text) Filter: ((id)::text = ANY ('{a4fe1cbb-8f3f-499e-9ae2-f4a5301b074e,1537e802-0458-4e75-ba53-47b4c60cc9b8,7e7aa6c8-c2dd-4373-b793-70440b940c29,fb5b2c90-a7cc-4734-b0e1-518f08e8972e,b8929b19-cfa7-4581-8867-41d091ca3ade,8e4559af-7755-4425-a409-e3cbe97669db,a50e4279-5f52-4e36-b2ed-b1de6c5086e4,ac94281e-6549-40db-bb7b-664c07bdb84e,80d6a2b8-9268-446a-bb0c-b787f49b262c,fe473ca5-d96f-49fb-959e-b70cf15919a9,291f2271-9645-4f0f-830b-9190786d78d3,92c9d569-fa32-4cca-b18e-8f69a0802a3c,f7dab2be-5b8e-4dfd-be16-af2bd2d6c6d6,87e1890e-9a54-4f85-8340-d341f9b9a1f3,16032ca6-3009-48eb-be54-6b32e2dce11f,8c410ba7-852c-4510-ad61-abb6e4e62706,cfdaacca-c6ca-431f-ab9c-fce60df46ee1,c7aa7e85-0076-4b1f-a2b6-8d49d67eefc6,a4ea061c-ac20-4515-b069-59ce8c8a0b1b,624826b0-1c73-4d73-a46a-84df45336b3b,9853c273-844c-4fd7-a64a-e27b7d44f577,c663f5b3-6a17-4863-aa8f-4d1c8a85200c,e0d499cf-9b41-4da3-b501-e29cb1dcc0a3}'::text[])) Rows Removed by Filter: 328 Heap Fetches: 0 -> Hash (cost=58.10..58.10 rows=655 width=79) (actual time=0.324..0.325 rows=655 loops=1) Buckets: 1024 Batches: 1 Memory Usage: 80kB -> Index Only Scan using ccs_xdevice_parent_asset_pseudo_level_idx on cssxxxaaaset facility_location (cost=0.28..58.10 rows=655 width=79) (actual time=0.030..0.122 rows=655 loops=1) Heap Fetches: 0 -> Index Scan using npspppordedddd_target_schedule_maintain_plan_status_idx on npspppordedddd (cost=0.56..4877.55 rows=359 width=726) (actual time=3.047..6.593 rows=1048 loops=23) Index Cond: (((target_code)::text = (cssxxxaaaset.id)::text) AND ((maintain_type)::text = 'DS'::text) AND (plan_time >= '2024-02-03 00:00:00'::timestamp without time zone) AND (plan_time <= '2024-03-03 00:00:00'::timestamp without time zone)) -> Index Scan using npsmanscheeee_pkey on npsmanscheeee (cost=0.42..4.44 rows=1 width=303) (actual time=0.004..0.004 rows=1 loops=24104) Index Cond: ((id)::text = (npspppordedddd.schedule_id)::text) -> Hash (cost=82.76..82.76 rows=493 width=1028) (actual time=1.066..1.067 rows=747 loops=1) Buckets: 1024 Batches: 1 Memory Usage: 149kB -> Seq Scan on npsmainfssdx (cost=0.00..82.76 rows=493 width=1028) (actual time=0.020..0.784 rows=747 loops=1) Filter: (((target_type)::text = 'device'::text) AND ((tenant_code)::text = ANY ('{_ADMIN,_NA_,59052881-df03-4c49-923f-6382b8ed138b,573bcfa0-239a-4f0b-9105-b906db4895ec,3264de0d-7262-4dce-8178-0112c0a1603b,0e0208f8-8763-4735-bb2c-ea5ef8bda742,f333dba3-7f08-4276-8bcb-a99c6ecc5120,4b04078d-8265-400d-a27d-e40d0f1646b4,52e44df1-be04-4bfc-bd5f-f37dbb60252c,6088e13a-28de-4b88-908f-247889e3d068,213fe35d-ddce-4876-a91c-f8c292713a20,e5fa7501-b3cb-4c59-bb78-6f8e5cdad19b,acda7d56-6d0f-4302-b7f4-292fb71173ba,ad13e67e-8bdc-45b1-96f4-f156c9c7dfb0,8b240bd1-bc1d-4dcf-87de-4e8e3e74d004,69869877-f9ce-4e85-a73b-edefe702bb6c,0ad151a9-bec9-4e23-953b-9f4c2e7be9a1}'::text[]))) Rows Removed by Filter: 46 -> Index Scan using npsmainproxxss_pkey on npsmainproxxss (cost=0.42..4.44 rows=1 width=515) (actual time=0.004..0.004 rows=1 loops=24104) Index Cond: ((id)::text = (npsmanscheeee.maintain_project_id)::text) -> Hash (cost=3.60..3.60 rows=60 width=367) (actual time=0.034..0.035 rows=0 loops=1) Buckets: 1024 Batches: 1 Memory Usage: 8kB -> Seq Scan on cssdsewff (cost=0.00..3.60 rows=60 width=367) (actual time=0.013..0.031 rows=60 loops=1) -> Hash (cost=1.19..1.19 rows=19 width=146) (actual time=0.014..0.015 rows=19 loops=1) Buckets: 1024 Batches: 1 Memory Usage: 10kB -> Seq Scan on ccssdeewqnn (cost=0.00..1.19 rows=19 width=146) (actual time=0.007..0.009 rows=19 loops=1) Planning Time: 3.460 ms Execution Time: 788.862 ms
可以看到 postgresql 这条分页SQL语句从 limit 改成 row_number 函数实现分页功能后，速度也能从原来跑不出结果降至 788ms 执行完成，本条SQL到此已经优化完毕。
最后问题：那为什么在Kingbase数据库上原SQL使用 limit 会慢？改成 rownum 后速度能秒出，通常情况下来说 limit 是PG提供原生的语法，性能应该更好才是？
解答：是因为在PostgreSQL中，LIMIT子句本身不直接与索引类型相关联，而是用于指定返回的记录数。然而，当LIMIT与ORDER BY结合使用时，PostgreSQL的查询优化器可能会利用B+树索引来加速查询。
这是因为B+树索引能够有效地支持有序数据的检索，使得数据库能够快速地定位到需要的记录而不必扫描整个表或索引。
然而需要通过索引进行排序的话，必然要通过 Index Scan 或者 Index Only Scan 扫描才可以对数据进行升序或者降序排序，而位图索引是不支持对数据进行排序功能的。
所以为什么一开始SQL会使用 Index Scan 和 Index Only Scan 而不使用 Bitmap Index Scan + BitmapOr 的查询策略。
各位读者以后在kingbase数据库进行业务开发，如果需要谓词过滤条件中有 or ，排序限制条件中有 order by + limit 的需求，尽量对业务SQL进行评估，从而选择使用 rownum 还是 limit 语句来进行限制数据。
如果在postgresql 进行开发的话遇到这种需求（pg不支持rownum写法），还需要在外面再包一层查询，使用 row_number() over() 窗口函数来进行限制即可。
