Mostafa Mokhtar created IMPALA-5851:
---------------------------------------
Summary: Estimate number of rows for sum_init_zero scans should
be number of files not table cardinality
Key: IMPALA-5851
URL: https://issues.apache.org/jira/browse/IMPALA-5851
Project: IMPALA
Issue Type: Bug
Components: Frontend
Reporter: Mostafa Mokhtar
Priority: Minor
IMPALA-5036 introduced an optimization to use the data stored in the Parquet
RowGroup.num_rows field for count(*) queries.
The estimate cardinality for the scan is the number of rows in the base table
opposed to number of files or row groups.
{code}
+-------------------------------------------------------------------------------+
| Explain String
|
+-------------------------------------------------------------------------------+
| Max Per-Host Resource Reservation: Memory=0B
|
| Per-Host Resource Estimates: Memory=108.00MB
|
|
|
| F01:PLAN FRAGMENT [UNPARTITIONED] hosts=1 instances=1
|
| | Per-Host Resources: mem-estimate=10.00MB mem-reservation=0B
|
| PLAN-ROOT SINK
|
| | mem-estimate=0B mem-reservation=0B
|
| |
|
| 03:AGGREGATE [FINALIZE]
|
| | output: count:merge(*)
|
| | mem-estimate=10.00MB mem-reservation=0B spill-buffer=2.00MB
|
| | tuple-ids=1 row-size=8B cardinality=1
|
| |
|
| 02:EXCHANGE [UNPARTITIONED]
|
| | mem-estimate=0B mem-reservation=0B
|
| | tuple-ids=1 row-size=8B cardinality=1
|
| |
|
| F00:PLAN FRAGMENT [RANDOM] hosts=130 instances=130
|
| Per-Host Resources: mem-estimate=98.00MB mem-reservation=0B
|
| 01:AGGREGATE
|
| | output: sum_init_zero(tpch_30000_parquet.lineitem.parquet-stats: num_rows)
|
| | mem-estimate=10.00MB mem-reservation=0B spill-buffer=2.00MB
|
| | tuple-ids=1 row-size=8B cardinality=1
|
| |
|
| 00:SCAN HDFS [tpch_30000_parquet.lineitem, RANDOM]
|
| partitions=2526/2526 files=28976 size=6.89TB
|
| stats-rows=179999978268 extrapolated-rows=disabled
|
| table stats: rows=179999978268 size=unavailable
|
| column stats: all
|
| mem-estimate=88.00MB mem-reservation=0B
|
| tuple-ids=0 row-size=8B cardinality=179999978268
|
+-------------------------------------------------------------------------------+
{code}
{code}
+--------------+--------+----------+----------+--------+------------+-----------+---------------+-----------------------------+
| Operator | #Hosts | Avg Time | Max Time | #Rows | Est. #Rows | Peak Mem
| Est. Peak Mem | Detail |
+--------------+--------+----------+----------+--------+------------+-----------+---------------+-----------------------------+
| 03:AGGREGATE | 1 | 1.28ms | 1.28ms | 1 | 1 | 532.00 KB
| 10.00 MB | FINALIZE |
| 02:EXCHANGE | 1 | 2.56s | 2.56s | 129 | 1 | 0 B
| 0 B | UNPARTITIONED |
| 01:AGGREGATE | 129 | 4.89ms | 62.84ms | 129 | 1 | 20.00 KB
| 10.00 MB | |
| 00:SCAN HDFS | 129 | 62.44ms | 341.03ms | 28.98K | 180.00B | 1.75 MB
| 88.00 MB | tpch_30000_parquet.lineitem |
+--------------+--------+----------+----------+--------+------------+-----------+---------------+-----------------------------+
{code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
