Merge pull request #66 from ClickHouse/pg-update

Author: rschu1ze

postgresql/results/_index_usage/_m6i.8xlarge_bluesky_1000m_lz4.index_usage

@@ -1,33 +1,33 @@
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q1:
 
- Sort  (cost=69519427.00..69519427.03 rows=13 width=40)
+ Sort  (cost=69172569.47..69172569.50 rows=12 width=40)
    Sort Key: (count(*)) DESC
-   ->  Finalize GroupAggregate  (cost=69519423.40..69519426.75 rows=13 width=40)
+   ->  Finalize GroupAggregate  (cost=69172566.16..69172569.26 rows=12 width=40)
          Group Key: (((data -> 'commit'::text) ->> 'collection'::text))
-         ->  Gather Merge  (cost=69519423.40..69519426.43 rows=26 width=40)
+         ->  Gather Merge  (cost=69172566.16..69172568.96 rows=24 width=40)
                Workers Planned: 2
-               ->  Sort  (cost=69518423.37..69518423.40 rows=13 width=40)
+               ->  Sort  (cost=69171566.13..69171566.16 rows=12 width=40)
                      Sort Key: (((data -> 'commit'::text) ->> 'collection'::text))
-                     ->  Partial HashAggregate  (cost=69518422.94..69518423.13 rows=13 width=40)
+                     ->  Partial HashAggregate  (cost=69171565.74..69171565.92 rows=12 width=40)
                            Group Key: ((data -> 'commit'::text) ->> 'collection'::text)
-                           ->  Parallel Seq Scan on bluesky  (cost=0.00..67846651.20 rows=334354347 width=32)
+                           ->  Parallel Seq Scan on bluesky  (cost=0.00..67495020.80 rows=335308987 width=32)
  JIT:
    Functions: 7
    Options: Inlining true, Optimization true, Expressions true, Deforming true
 
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q2:
 
- Sort  (cost=534467372.01..534467372.04 rows=13 width=48)
+ Sort  (cost=531988506.66..531988506.69 rows=12 width=48)
    Sort Key: (count(*)) DESC
-   ->  GroupAggregate  (cost=438967879.95..534467371.77 rows=13 width=48)
+   ->  GroupAggregate  (cost=436041090.60..531988506.44 rows=12 width=48)
          Group Key: (((data -> 'commit'::text) ->> 'collection'::text))
-         ->  Gather Merge  (cost=438967879.95..526916003.64 rows=755136794 width=569)
+         ->  Gather Merge  (cost=436041090.60..524401719.92 rows=758678634 width=566)
                Workers Planned: 2
-               ->  Sort  (cost=438966879.92..439753480.75 rows=314640331 width=569)
+               ->  Sort  (cost=436040090.57..436830380.82 rows=316116098 width=566)
                      Sort Key: (((data -> 'commit'::text) ->> 'collection'::text)), ((data ->> 'did'::text))
-                     ->  Parallel Seq Scan on bluesky  (cost=0.00..71927510.45 rows=314640331 width=569)
+                     ->  Parallel Seq Scan on bluesky  (cost=0.00..71590418.69 rows=316116098 width=566)
                            Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text))
  JIT:
    Functions: 8
@@ -36,15 +36,15 @@ Index usage for query Q2:
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q3:
 
- Finalize GroupAggregate  (cost=134562855.16..206828438.13 rows=473420435 width=72)
+ Finalize GroupAggregate  (cost=134041870.72..206094379.64 rows=472024562 width=72)
    Group Key: (EXTRACT(hour FROM to_timestamp(((((data ->> 'time_us'::text))::bigint / 1000000))::double precision))), (((data -> 'commit'::text) ->> 'collection'::text))
-   ->  Gather Merge  (cost=134562855.16..188483396.27 rows=394517030 width=72)
+   ->  Gather Merge  (cost=134041870.72..187803427.86 rows=393353802 width=72)
          Workers Planned: 2
-         ->  Partial GroupAggregate  (cost=134561855.14..142945342.03 rows=197258515 width=72)
+         ->  Partial GroupAggregate  (cost=134040870.70..142399638.99 rows=196676901 width=72)
                Group Key: (EXTRACT(hour FROM to_timestamp(((((data ->> 'time_us'::text))::bigint / 1000000))::double precision))), (((data -> 'commit'::text) ->> 'collection'::text))
-               ->  Sort  (cost=134561855.14..135055001.43 rows=197258515 width=64)
+               ->  Sort  (cost=134040870.70..134532562.95 rows=196676901 width=64)
                      Sort Key: (EXTRACT(hour FROM to_timestamp(((((data ->> 'time_us'::text))::bigint / 1000000))::double precision))), (((data -> 'commit'::text) ->> 'collection'::text))
-                     ->  Parallel Seq Scan on bluesky  (cost=0.00..77718225.92 rows=197258515 width=64)
+                     ->  Parallel Seq Scan on bluesky  (cost=0.00..77369022.11 rows=196676901 width=64)
                            Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text) AND (((data -> 'commit'::text) ->> 'collection'::text) = ANY ('{app.bsky.feed.post,app.bsky.feed.repost,app.bsky.feed.like}'::text[])))
  JIT:
    Functions: 11
@@ -53,39 +53,42 @@ Index usage for query Q3:
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q4:
 
- Limit  (cost=116709176.13..116709176.13 rows=3 width=40)
-   ->  Sort  (cost=116709176.13..116885878.13 rows=70680802 width=40)
+ Limit  (cost=94926225.45..94926225.46 rows=3 width=40)
+   ->  Sort  (cost=94926225.45..94926749.83 rows=209751 width=40)
          Sort Key: (min(('1970-01-01 00:00:00+00'::timestamp with time zone + ('00:00:00.000001'::interval * (((data ->> 'time_us'::text))::bigint)::double precision))))
-         ->  Finalize GroupAggregate  (cost=106788266.40..115795640.01 rows=70680802 width=40)
+         ->  Finalize GroupAggregate  (cost=94869849.75..94923514.46 rows=209751 width=40)
                Group Key: ((data ->> 'did'::text))
-               ->  Gather Merge  (cost=106788266.40..114617626.65 rows=58900668 width=40)
+               ->  Gather Merge  (cost=94869849.75..94918795.06 rows=419502 width=40)
                      Workers Planned: 2
-                     ->  Partial GroupAggregate  (cost=106787266.37..107818028.06 rows=29450334 width=40)
-                           Group Key: ((data ->> 'did'::text))
-                           ->  Sort  (cost=106787266.37..106860892.21 rows=29450334 width=569)
-                                 Sort Key: ((data ->> 'did'::text))
-                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..72935592.23 rows=29450334 width=569)
+                     ->  Sort  (cost=94868849.72..94869374.10 rows=209751 width=40)
+                           Sort Key: ((data ->> 'did'::text))
+                           ->  Partial HashAggregate  (cost=90626411.32..94844573.00 rows=209751 width=40)
+                                 Group Key: (data ->> 'did'::text)
+                                 Planned Partitions: 32
+                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..72597573.04 rows=29166978 width=566)
                                        Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text) AND (((data -> 'commit'::text) ->> 'collection'::text) = 'app.bsky.feed.post'::text))
  JIT:
-   Functions: 13
+   Functions: 10
    Options: Inlining true, Optimization true, Expressions true, Deforming true
 
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q5:
 
- Limit  (cost=117901914.66..117901914.66 rows=3 width=64)
-   ->  Sort  (cost=117901914.66..118078616.66 rows=70680802 width=64)
+ Limit  (cost=95439986.96..95439986.96 rows=3 width=64)
+   ->  Sort  (cost=95439986.96..95440511.33 rows=209751 width=64)
          Sort Key: ((EXTRACT(epoch FROM (max(('1970-01-01 00:00:00+00'::timestamp with time zone + ('00:00:00.000001'::interval * (((data ->> 'time_us'::text))::bigint)::double precision))) - min(('1970-01-01 00:00:00+00'::timestamp with time zone + ('00:00:00.000001'::interval * (((data ->> 'time_us'::text))::bigint)::double precision))))) * '1000'::numeric)) DESC
-         ->  Finalize GroupAggregate  (cost=106788266.40..116988378.54 rows=70680802 width=64)
+         ->  Finalize GroupAggregate  (cost=95380989.36..95437275.96 rows=209751 width=64)
                Group Key: ((data ->> 'did'::text))
-               ->  Gather Merge  (cost=106788266.40..115133007.49 rows=58900668 width=48)
+               ->  Gather Merge  (cost=95380989.36..95429934.68 rows=419502 width=48)
                      Workers Planned: 2
-                     ->  Partial GroupAggregate  (cost=106787266.37..108333408.91 rows=29450334 width=48)
-                           Group Key: ((data ->> 'did'::text))
-                           ->  Sort  (cost=106787266.37..106860892.21 rows=29450334 width=569)
-                                 Sort Key: ((data ->> 'did'::text))
-                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..72935592.23 rows=29450334 width=569)
+                     ->  Sort  (cost=95379989.34..95380513.71 rows=209751 width=48)
+                           Sort Key: ((data ->> 'did'::text))
+                           ->  Partial HashAggregate  (cost=91136833.44..95354995.11 rows=209751 width=48)
+                                 Group Key: (data ->> 'did'::text)
+                                 Planned Partitions: 32
+                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..72597573.04 rows=29166978 width=566)
                                        Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text) AND (((data -> 'commit'::text) ->> 'collection'::text) = 'app.bsky.feed.post'::text))
  JIT:
-   Functions: 13
-   Options: Inlining true, Optimization true, Expressions true, Deforming true
+   Functions: 10
+   Options: Inlining true, Optimization true, Expressions true, Deforming true
+

postgresql/results/_index_usage/_m6i.8xlarge_bluesky_1000m_pglz.index_usage

@@ -1,33 +1,33 @@
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q1:
 
- Sort  (cost=69897029.86..69897029.89 rows=13 width=40)
+ Sort  (cost=69151118.72..69151118.75 rows=13 width=40)
    Sort Key: (count(*)) DESC
-   ->  Finalize GroupAggregate  (cost=69897026.26..69897029.62 rows=13 width=40)
+   ->  Finalize GroupAggregate  (cost=69151115.12..69151118.48 rows=13 width=40)
          Group Key: (((data -> 'commit'::text) ->> 'collection'::text))
-         ->  Gather Merge  (cost=69897026.26..69897029.29 rows=26 width=40)
+         ->  Gather Merge  (cost=69151115.12..69151118.16 rows=26 width=40)
                Workers Planned: 2
-               ->  Sort  (cost=69896026.24..69896026.27 rows=13 width=40)
+               ->  Sort  (cost=69150115.10..69150115.13 rows=13 width=40)
                      Sort Key: (((data -> 'commit'::text) ->> 'collection'::text))
-                     ->  Partial HashAggregate  (cost=69896025.80..69896025.99 rows=13 width=40)
+                     ->  Partial HashAggregate  (cost=69150114.66..69150114.86 rows=13 width=40)
                            Group Key: ((data -> 'commit'::text) ->> 'collection'::text)
-                           ->  Parallel Seq Scan on bluesky  (cost=0.00..68213928.60 rows=336419440 width=32)
+                           ->  Parallel Seq Scan on bluesky  (cost=0.00..67475856.00 rows=334851733 width=32)
  JIT:
    Functions: 7
    Options: Inlining true, Optimization true, Expressions true, Deforming true
 
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q2:
 
- Sort  (cost=533218266.90..533218266.93 rows=13 width=48)
+ Sort  (cost=532206068.60..532206068.63 rows=13 width=48)
    Sort Key: (count(*)) DESC
-   ->  GroupAggregate  (cost=437166993.49..533218266.66 rows=13 width=48)
+   ->  GroupAggregate  (cost=436209018.58..532206068.36 rows=13 width=48)
          Group Key: (((data -> 'commit'::text) ->> 'collection'::text))
-         ->  Gather Merge  (cost=437166993.49..525623267.88 rows=759499858 width=568)
+         ->  Gather Merge  (cost=436209018.58..524615357.15 rows=759071101 width=568)
                Workers Planned: 2
-               ->  Sort  (cost=437165993.47..437957139.15 rows=316458274 width=568)
+               ->  Sort  (cost=436208018.55..436998717.62 rows=316279625 width=568)
                      Sort Key: (((data -> 'commit'::text) ->> 'collection'::text)), ((data ->> 'did'::text))
-                     ->  Parallel Seq Scan on bluesky  (cost=0.00..72319365.77 rows=316458274 width=568)
+                     ->  Parallel Seq Scan on bluesky  (cost=0.00..71568642.12 rows=316279625 width=568)
                            Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text))
  JIT:
    Functions: 8
@@ -36,15 +36,15 @@ Index usage for query Q2:
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q3:
 
- Finalize GroupAggregate  (cost=134776076.46..206825424.55 rows=472003856 width=72)
+ Finalize GroupAggregate  (cost=134473230.64..207053525.23 rows=475482150 width=72)
    Group Key: (EXTRACT(hour FROM to_timestamp(((((data ->> 'time_us'::text))::bigint / 1000000))::double precision))), (((data -> 'commit'::text) ->> 'collection'::text))
-   ->  Gather Merge  (cost=134776076.46..188535275.14 rows=393336546 width=72)
+   ->  Gather Merge  (cost=134473230.64..188628591.92 rows=396235124 width=72)
          Workers Planned: 2
-         ->  Partial GroupAggregate  (cost=134775076.44..143133478.04 rows=196668273 width=72)
+         ->  Partial GroupAggregate  (cost=134472230.62..142892227.00 rows=198117562 width=72)
                Group Key: (EXTRACT(hour FROM to_timestamp(((((data ->> 'time_us'::text))::bigint / 1000000))::double precision))), (((data -> 'commit'::text) ->> 'collection'::text))
-               ->  Sort  (cost=134775076.44..135266747.12 rows=196668273 width=64)
+               ->  Sort  (cost=134472230.62..134967524.52 rows=198117562 width=64)
                      Sort Key: (EXTRACT(hour FROM to_timestamp(((((data ->> 'time_us'::text))::bigint / 1000000))::double precision))), (((data -> 'commit'::text) ->> 'collection'::text))
-                     ->  Parallel Seq Scan on bluesky  (cost=0.00..78105780.64 rows=196668273 width=64)
+                     ->  Parallel Seq Scan on bluesky  (cost=0.00..77374841.81 rows=198117562 width=64)
                            Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text) AND (((data -> 'commit'::text) ->> 'collection'::text) = ANY ('{app.bsky.feed.post,app.bsky.feed.repost,app.bsky.feed.like}'::text[])))
  JIT:
    Functions: 11
@@ -53,39 +53,42 @@ Index usage for query Q3:
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q4:
 
- Limit  (cost=118661659.29..118661659.29 rows=3 width=40)
-   ->  Sort  (cost=118661659.29..118846281.04 rows=73848702 width=40)
+ Limit  (cost=95640269.60..95640269.61 rows=3 width=40)
+   ->  Sort  (cost=95640269.60..95640809.49 rows=215958 width=40)
          Sort Key: (min(('1970-01-01 00:00:00+00'::timestamp with time zone + ('00:00:00.000001'::interval * (((data ->> 'time_us'::text))::bigint)::double precision))))
-         ->  Finalize GroupAggregate  (cost=108296096.38..117707178.66 rows=73848702 width=40)
+         ->  Finalize GroupAggregate  (cost=95582225.61..95637478.38 rows=215958 width=40)
                Group Key: ((data ->> 'did'::text))
-               ->  Gather Merge  (cost=108296096.38..116476366.96 rows=61540584 width=40)
+               ->  Gather Merge  (cost=95582225.61..95632619.33 rows=431916 width=40)
                      Workers Planned: 2
-                     ->  Partial GroupAggregate  (cost=108295096.35..109372056.57 rows=30770292 width=40)
-                           Group Key: ((data ->> 'did'::text))
-                           ->  Sort  (cost=108295096.35..108372022.08 rows=30770292 width=568)
-                                 Sort Key: ((data ->> 'did'::text))
-                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..73337145.93 rows=30770292 width=568)
+                     ->  Sort  (cost=95581225.59..95581765.48 rows=215958 width=40)
+                           Sort Key: ((data ->> 'did'::text))
+                           ->  Partial HashAggregate  (cost=91198626.16..95556183.28 rows=215958 width=40)
+                                 Group Key: (data ->> 'did'::text)
+                                 Planned Partitions: 32
+                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..72573959.27 rows=30130907 width=568)
                                        Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text) AND (((data -> 'commit'::text) ->> 'collection'::text) = 'app.bsky.feed.post'::text))
  JIT:
-   Functions: 13
+   Functions: 10
    Options: Inlining true, Optimization true, Expressions true, Deforming true
 
 ------------------------------------------------------------------------------------------------------------------------
 Index usage for query Q5:
 
- Limit  (cost=119907856.12..119907856.13 rows=3 width=64)
-   ->  Sort  (cost=119907856.12..120092477.88 rows=73848702 width=64)
+ Limit  (cost=96170998.45..96170998.45 rows=3 width=64)
+   ->  Sort  (cost=96170998.45..96171538.34 rows=215958 width=64)
          Sort Key: ((EXTRACT(epoch FROM (max(('1970-01-01 00:00:00+00'::timestamp with time zone + ('00:00:00.000001'::interval * (((data ->> 'time_us'::text))::bigint)::double precision))) - min(('1970-01-01 00:00:00+00'::timestamp with time zone + ('00:00:00.000001'::interval * (((data ->> 'time_us'::text))::bigint)::double precision))))) * '1000'::numeric)) DESC
-         ->  Finalize GroupAggregate  (cost=108296096.38..118953375.49 rows=73848702 width=64)
+         ->  Finalize GroupAggregate  (cost=96110254.98..96168207.23 rows=215958 width=64)
                Group Key: ((data ->> 'did'::text))
-               ->  Gather Merge  (cost=108296096.38..117014847.07 rows=61540584 width=48)
+               ->  Gather Merge  (cost=96110254.98..96160648.70 rows=431916 width=48)
                      Workers Planned: 2
-                     ->  Partial GroupAggregate  (cost=108295096.35..109910536.68 rows=30770292 width=48)
-                           Group Key: ((data ->> 'did'::text))
-                           ->  Sort  (cost=108295096.35..108372022.08 rows=30770292 width=568)
-                                 Sort Key: ((data ->> 'did'::text))
-                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..73337145.93 rows=30770292 width=568)
+                     ->  Sort  (cost=96109254.96..96109794.85 rows=215958 width=48)
+                           Sort Key: ((data ->> 'did'::text))
+                           ->  Partial HashAggregate  (cost=91725917.03..96083474.16 rows=215958 width=48)
+                                 Group Key: (data ->> 'did'::text)
+                                 Planned Partitions: 32
+                                 ->  Parallel Seq Scan on bluesky  (cost=0.00..72573959.27 rows=30130907 width=568)
                                        Filter: (((data ->> 'kind'::text) = 'commit'::text) AND (((data -> 'commit'::text) ->> 'operation'::text) = 'create'::text) AND (((data -> 'commit'::text) ->> 'collection'::text) = 'app.bsky.feed.post'::text))
  JIT:
-   Functions: 13
-   Options: Inlining true, Optimization true, Expressions true, Deforming true
+   Functions: 10
+   Options: Inlining true, Optimization true, Expressions true, Deforming true
+