alamb commented on issue #18070:
URL: https://github.com/apache/datafusion/issues/18070#issuecomment-3418219330

   > Could you check the size of the inputs for the NLJoin? If this is NLJoin 
related it could be similar to 
[#17547](https://github.com/apache/datafusion/issues/17547) and 
[#17488](https://github.com/apache/datafusion/issues/17488) .
   
   I ran `EXPLAIN ANAYZE`
   
   The relevant part of the plan I think is like this (tiny left input - 1 row, 
giant right input 22M rows)
   ```
   NLJoin
      ProjectionExec (1 row)
      CoalesceBatchesExec (21917655 rows)
   ```
   
   The full plan is here:
   
   <details><summary>Full Output</summary>
   <p>
   
   ```sql
   
+-------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 -----+
   | plan_type         | plan                                                   
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   
+-------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 -----+
   | Plan with Metrics | ProjectionExec: expr=[count(Int64(1))@0 as count(*)], 
metrics=[output_rows=1, elapsed_compute=250ns]                                  
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                       
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |   AggregateExec: mode=Final, gby=[], 
aggr=[count(Int64(1))], metrics=[output_rows=1, elapsed_compute=431.127µs]      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                       
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
       |
   |                   |     CoalescePartitionsExec, metrics=[output_rows=16, 
elapsed_compute=12.875µs]                                                       
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                       
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
       |
   |                   |       AggregateExec: mode=Partial, gby=[], 
aggr=[count(Int64(1))], metrics=[output_rows=16, elapsed_compute=20.151647ms]   
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                  
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |         ProjectionExec: expr=[], 
metrics=[output_rows=21018273, elapsed_compute=837.882µs]                       
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                           
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
       |
   |                   |           CoalesceBatchesExec: target_batch_size=8192, 
metrics=[output_rows=21018273, elapsed_compute=17.333442058s]                   
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |             FilterExec: 
array_has_any(fsq_category_ids@0, category_ids@1), 
metrics=[output_rows=21018273, elapsed_compute=173.866690988s]                  
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                  
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |               ProjectionExec: expr=[fsq_category_ids@1 
as fsq_category_ids, category_ids@0 as category_ids], 
metrics=[output_rows=21917655, elapsed_compute=2.68371ms]                       
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |                 NestedLoopJoinExec: join_type=Right, 
metrics=[output_rows=21917655, elapsed_compute=35.600322924s, 
build_input_batches=1, build_input_rows=1, input_batches=2356, 
input_rows=21917655, output_batches=4696, build_mem_used=55068, 
build_time=7.882335ms, join_time=35.590911587s]                                 
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
           
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |                   ProjectionExec: 
expr=[array_agg(categories_raw.category_id)@0 as category_ids], 
metrics=[output_rows=1, elapsed_compute=43.375µs]                               
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                          
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
       |
   |                   |                     GlobalLimitExec: skip=0, 
fetch=500, metrics=[output_rows=1, elapsed_compute=42ns]                        
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |                       AggregateExec: mode=Single, 
gby=[], aggr=[array_agg(categories_raw.category_id)], metrics=[output_rows=1, 
elapsed_compute=79.125µs]                                                       
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                            
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
       |
   |                   |                         DataSourceExec: file_groups={1 
group: [[release/dt=2025-09-09/categories/parquet/categories.zstd.parquet]]}, 
projection=[category_id], file_type=parquet, metrics=[output_rows=1245, 
elapsed_compute=1ns, batches_splitted=0, bytes_scanned=6295, 
file_open_errors=0, file_scan_errors=0, files_ranges_pruned_statistics=0, 
num_predicate_creation_errors=0, page_index_rows_matched=0, 
page_index_rows_pruned=0, predicate_evaluation_errors=0, 
pushdown_rows_matched=0, pushdown_rows_pruned=0, 
row_groups_matched_bloom_filter=0, row_groups_matched_statistics=0, 
row_groups_pruned_bloom_filter=0, row_groups_pruned_statistics=0, 
bloom_filter_eval_time=2ns, metadata_load_time=89.590751ms, 
page_index_eval_time=2ns, row_pushdown_eval_time=2ns, statistics_eval_time=2ns, 
time_elapsed_opening=90.102958ms, time_elapsed_processing=7.179959ms, 
time_elapsed_scanning_total=51.312167ms, 
time_elapsed_scanning_until_data=51.186792ms]                                   
  
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |                   CoalesceBatchesExec: 
target_batch_size=8192, metrics=[output_rows=21917655, 
elapsed_compute=80.199596ms]                                                    
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                               
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |                     FilterExec: CAST(date_refreshed@0 
AS Date32) >= 2024-10-18, projection=[fsq_category_ids@1], 
metrics=[output_rows=21917655, elapsed_compute=1.918010644s]                    
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                            
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   |                   |                       DataSourceExec: file_groups={16 
groups: 
[[release/dt=2025-09-09/places/parquet/places-00000.zstd.parquet:0..115950054, 
release/dt=2025-09-09/places/parquet/places-00001.zstd.parquet:0..119625610, 
release/dt=2025-09-09/places/parquet/places-00002.zstd.parquet:0..111979679, 
release/dt=2025-09-09/places/parquet/places-00003.zstd.parquet:0..99590199, 
release/dt=2025-09-09/places/parquet/places-00004.zstd.parquet:0..114422801, 
...], 
[release/dt=2025-09-09/places/parquet/places-00006.zstd.parquet:15283909..123129325,
 release/dt=2025-09-09/places/parquet/places-00007.zstd.parquet:0..109503606, 
release/dt=2025-09-09/places/parquet/places-00008.zstd.parquet:0..116880554, 
release/dt=2025-09-09/places/parquet/places-00009.zstd.parquet:0..107480580, 
release/dt=2025-09-09/places/parquet/places-00010.zstd.parquet:0..112300670, 
...], 
[release/dt=2025-09-09/places/parquet/places-00012.zstd.parquet:37168422..107444227,
 release/dt=2025-09-09/places/parque
 t/places-00013.zstd.parquet:0..110757380, 
release/dt=2025-09-09/places/parquet/places-00014.zstd.parquet:0..122112592, 
release/dt=2025-09-09/places/parquet/places-00015.zstd.parquet:0..119978222, 
release/dt=2025-09-09/places/parquet/places-00016.zstd.parquet:0..128256136, 
...], 
[release/dt=2025-09-09/places/parquet/places-00018.zstd.parquet:44848422..120106873,
 release/dt=2025-09-09/places/parquet/places-00019.zstd.parquet:0..118004710, 
release/dt=2025-09-09/places/parquet/places-00020.zstd.parquet:0..118719840, 
release/dt=2025-09-09/places/parquet/places-00021.zstd.parquet:0..118914980, 
release/dt=2025-09-09/places/parquet/places-00022.zstd.parquet:0..120055444, 
...], 
[release/dt=2025-09-09/places/parquet/places-00024.zstd.parquet:22110584..112529288,
 release/dt=2025-09-09/places/parquet/places-00025.zstd.parquet:0..117357063, 
release/dt=2025-09-09/places/parquet/places-00026.zstd.parquet:0..101323034, 
release/dt=2025-09-09/places/parquet/places-00027.zstd.parquet:0..113101190, rel
 ease/dt=2025-09-09/places/parquet/places-00028.zstd.parquet:0..113015606, 
...], ...]}, projection=[date_refreshed, fsq_category_ids], file_type=parquet, 
predicate=CAST(date_refreshed@0 AS Date32) >= 2024-10-18, 
metrics=[output_rows=105292526, elapsed_compute=16ns, batches_splitted=0, 
bytes_scanned=327606810, file_open_errors=0, file_scan_errors=0, 
files_ranges_pruned_statistics=0, num_predicate_creation_errors=0, 
page_index_rows_matched=0, page_index_rows_pruned=0, 
predicate_evaluation_errors=0, pushdown_rows_matched=0, pushdown_rows_pruned=0, 
row_groups_matched_bloom_filter=0, row_groups_matched_statistics=0, 
row_groups_pruned_bloom_filter=0, row_groups_pruned_statistics=0, 
bloom_filter_eval_time=230ns, metadata_load_time=41.780586406s, 
page_index_eval_time=15.576µs, row_pushdown_eval_time=230ns, 
statistics_eval_time=230ns, time_elapsed_opening=2.202947921s, 
time_elapsed_processing=3.475536283s, 
time_elapsed_scanning_total=695.001470448s, 
time_elapsed_scanning_until_data=64.328045
 286s] |
   |                   |                                                        
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                                                                
                                      
      |
   
+-------------------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 -----+
   1 row(s) fetched.
   Elapsed 58.123 seconds.
   ```
   
   </p>
   </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: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to