This is an automated email from the ASF dual-hosted git repository.
morningman pushed a commit to branch tpcds
in repository https://gitbox.apache.org/repos/asf/doris.git
commit d39a5e03e59fca99ef2f3d9c6f946b8cf860ee06
Author: jakevin <jakevin...@gmail.com>
AuthorDate: Thu Jan 18 17:18:41 2024 +0800
[fix](Nereids): fix AssertNumRows StatsCalculator (#30053)
---
.../doris/nereids/stats/StatsCalculator.java | 44 +++++++++++---
.../shape/query54.out | 59 +++++++++++++++++-
.../nereids_tpcds_shape_sf1000_p0/shape/query6.out | 68 ++++++++++-----------
.../noStatsRfPrune/query58.out | 21 +++----
.../no_stats_shape/query58.out | 21 +++----
.../rf_prune/query54.out | 71 ++++++++++++++++++++--
.../rf_prune/query6.out | 7 +--
.../nereids_tpcds_shape_sf100_p0/shape/query54.out | 71 ++++++++++++++++++++--
.../nereids_tpcds_shape_sf100_p0/shape/query6.out | 9 ++-
9 files changed, 284 insertions(+), 87 deletions(-)
diff --git
a/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/StatsCalculator.java
b/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/StatsCalculator.java
index 14ead9990aa..76587bcffb5 100644
---
a/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/StatsCalculator.java
+++
b/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/StatsCalculator.java
@@ -26,6 +26,7 @@ import org.apache.doris.nereids.CascadesContext;
import org.apache.doris.nereids.memo.Group;
import org.apache.doris.nereids.memo.GroupExpression;
import org.apache.doris.nereids.trees.expressions.Alias;
+import org.apache.doris.nereids.trees.expressions.AssertNumRowsElement;
import org.apache.doris.nereids.trees.expressions.CTEId;
import org.apache.doris.nereids.trees.expressions.Expression;
import org.apache.doris.nereids.trees.expressions.NamedExpression;
@@ -167,7 +168,7 @@ public class StatsCalculator extends
DefaultPlanVisitor<Statistics, Void> {
private CascadesContext cascadesContext;
private StatsCalculator(GroupExpression groupExpression, boolean
forbidUnknownColStats,
- Map<String, ColumnStatistic>
columnStatisticMap, boolean isPlayNereidsDump,
+ Map<String, ColumnStatistic> columnStatisticMap, boolean
isPlayNereidsDump,
Map<CTEId, Statistics> cteIdToStats, CascadesContext context) {
this.groupExpression = groupExpression;
this.forbidUnknownColStats = forbidUnknownColStats;
@@ -193,7 +194,7 @@ public class StatsCalculator extends
DefaultPlanVisitor<Statistics, Void> {
* estimate stats
*/
public static StatsCalculator estimate(GroupExpression groupExpression,
boolean forbidUnknownColStats,
- Map<String, ColumnStatistic>
columnStatisticMap, boolean isPlayNereidsDump,
+ Map<String, ColumnStatistic> columnStatisticMap, boolean
isPlayNereidsDump,
Map<CTEId, Statistics> cteIdToStats, CascadesContext context) {
StatsCalculator statsCalculator = new StatsCalculator(
groupExpression, forbidUnknownColStats, columnStatisticMap,
isPlayNereidsDump, cteIdToStats, context);
@@ -369,7 +370,7 @@ public class StatsCalculator extends
DefaultPlanVisitor<Statistics, Void> {
@Override
public Statistics visitLogicalAssertNumRows(
LogicalAssertNumRows<? extends Plan> assertNumRows, Void context) {
- return
computeAssertNumRows(assertNumRows.getAssertNumRowsElement().getDesiredNumOfRows());
+ return computeAssertNumRows(assertNumRows.getAssertNumRowsElement());
}
@Override
@@ -533,7 +534,7 @@ public class StatsCalculator extends
DefaultPlanVisitor<Statistics, Void> {
@Override
public Statistics visitPhysicalAssertNumRows(PhysicalAssertNumRows<?
extends Plan> assertNumRows,
Void context) {
- return
computeAssertNumRows(assertNumRows.getAssertNumRowsElement().getDesiredNumOfRows());
+ return computeAssertNumRows(assertNumRows.getAssertNumRowsElement());
}
@Override
@@ -556,11 +557,34 @@ public class StatsCalculator extends
DefaultPlanVisitor<Statistics, Void> {
return computeGenerate(generate);
}
- private Statistics computeAssertNumRows(long desiredNumOfRows) {
+ private Statistics computeAssertNumRows(AssertNumRowsElement
assertNumRowsElement) {
Statistics statistics = groupExpression.childStatistics(0);
- statistics.withRowCountAndEnforceValid(Math.min(1,
statistics.getRowCount()));
- statistics = new
StatisticsBuilder(statistics).setWidthInJoinCluster(1).build();
- return statistics;
+ long newRowCount;
+ long rowCount = (long) statistics.getRowCount();
+ long desiredNumOfRows = assertNumRowsElement.getDesiredNumOfRows();
+ switch (assertNumRowsElement.getAssertion()) {
+ case EQ:
+ newRowCount = desiredNumOfRows;
+ break;
+ case GE:
+ newRowCount = statistics.getRowCount() >= desiredNumOfRows ?
rowCount : desiredNumOfRows;
+ break;
+ case GT:
+ newRowCount = statistics.getRowCount() > desiredNumOfRows ?
rowCount : desiredNumOfRows;
+ break;
+ case LE:
+ newRowCount = statistics.getRowCount() <= desiredNumOfRows ?
rowCount : desiredNumOfRows;
+ break;
+ case LT:
+ newRowCount = statistics.getRowCount() < desiredNumOfRows ?
rowCount : desiredNumOfRows;
+ break;
+ case NE:
+ return statistics;
+ default:
+ throw new IllegalArgumentException("Unknown assertion: " +
assertNumRowsElement.getAssertion());
+ }
+ Statistics newStatistics =
statistics.withRowCountAndEnforceValid(newRowCount);
+ return new
StatisticsBuilder(newStatistics).setWidthInJoinCluster(1).build();
}
private Statistics computeFilter(Filter filter) {
@@ -610,7 +634,7 @@ public class StatsCalculator extends
DefaultPlanVisitor<Statistics, Void> {
}
} else {
return
Env.getCurrentEnv().getStatisticsCache().getColumnStatistics(
- catalogId, dbId, table.getId(), colName);
+ catalogId, dbId, table.getId(), colName);
}
}
@@ -701,7 +725,7 @@ public class StatsCalculator extends
DefaultPlanVisitor<Statistics, Void> {
rowCount = rowCount * DEFAULT_COLUMN_NDV_RATIO;
} else {
rowCount = Math.min(rowCount,
partitionByKeyStats.stream().map(s -> s.ndv)
- .max(Double::compare).get() *
partitionTopN.getPartitionLimit());
+ .max(Double::compare).get() *
partitionTopN.getPartitionLimit());
}
} else {
rowCount = Math.min(rowCount, partitionTopN.getPartitionLimit());
diff --git
a/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query54.out
b/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query54.out
index a732fd81b3b..23f67fe4b3c 100644
--- a/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query54.out
+++ b/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query54.out
@@ -15,10 +15,62 @@ PhysicalResultSink
------------------------PhysicalProject
--------------------------hashJoin[INNER_JOIN]
hashCondition=((store_sales.ss_sold_date_sk = date_dim.d_date_sk))
otherCondition=() build RFs:RF6 ss_sold_date_sk->[d_date_sk]
----------------------------PhysicalProject
-------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) <= (d_month_seq + 3))
---------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) >= (d_month_seq + 1))
+------------------------------hashJoin[INNER_JOIN]
hashCondition=((my_customers.c_customer_sk = store_sales.ss_customer_sk))
otherCondition=() build RFs:RF6 c_customer_sk->[ss_customer_sk]
+--------------------------------PhysicalProject
+----------------------------------PhysicalOlapScan[store_sales] apply RFs: RF6
RF7
+--------------------------------PhysicalDistribute[DistributionSpecReplicated]
----------------------------------PhysicalProject
------------------------------------PhysicalOlapScan[date_dim] apply RFs: RF6
+----------------------------------------PhysicalProject
+------------------------------------------PhysicalOlapScan[customer_address]
apply RFs: RF3 RF4 RF5
+----------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------PhysicalProject
+--------------------------------------------hashAgg[GLOBAL]
+----------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------------hashAgg[LOCAL]
+--------------------------------------------------PhysicalProject
+----------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((customer.c_customer_sk = cs_or_ws_sales.customer_sk))
otherCondition=() build RFs:RF2 customer_sk->[c_customer_sk]
+------------------------------------------------------PhysicalProject
+--------------------------------------------------------PhysicalOlapScan[customer]
apply RFs: RF2
+------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------------------------PhysicalProject
+----------------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((cs_or_ws_sales.sold_date_sk = date_dim.d_date_sk))
otherCondition=() build RFs:RF1 d_date_sk->[cs_sold_date_sk,ws_sold_date_sk]
+------------------------------------------------------------PhysicalProject
+--------------------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((cs_or_ws_sales.item_sk = item.i_item_sk)) otherCondition=()
build RFs:RF0 i_item_sk->[cs_item_sk,ws_item_sk]
+----------------------------------------------------------------PhysicalUnion
+------------------------------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------------PhysicalOlapScan[catalog_sales]
apply RFs: RF0 RF1
+------------------------------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------------PhysicalOlapScan[web_sales]
apply RFs: RF0 RF1
+----------------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------------------------------------------PhysicalProject
+--------------------------------------------------------------------filter((item.i_category
= 'Music') and (item.i_class = 'country'))
+----------------------------------------------------------------------PhysicalOlapScan[item]
+------------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------filter((date_dim.d_moy
= 1) and (date_dim.d_year = 1999))
+------------------------------------------------------------------PhysicalOlapScan[date_dim]
+--------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+----------------------------------------PhysicalProject
+------------------------------------------PhysicalOlapScan[store]
+----------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------PhysicalProject
+--------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) <= (d_month_seq + 3))
+----------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq
as BIGINT) >= (d_month_seq + 1))
+------------------------------------PhysicalProject
+--------------------------------------PhysicalOlapScan[date_dim]
+------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------PhysicalAssertNumRows
+----------------------------------------PhysicalDistribute[DistributionSpecGather]
+------------------------------------------hashAgg[GLOBAL]
+--------------------------------------------PhysicalDistribute[DistributionSpecHash]
+----------------------------------------------hashAgg[LOCAL]
+------------------------------------------------PhysicalProject
+--------------------------------------------------filter((date_dim.d_moy = 1)
and (date_dim.d_year = 1999))
+----------------------------------------------------PhysicalOlapScan[date_dim]
+>>>>>>> 9c7d4b518e ([fix](Nereids): fix AssertNumRows StatsCalculator (#30053))
----------------------------------PhysicalDistribute[DistributionSpecReplicated]
------------------------------------PhysicalAssertNumRows
--------------------------------------PhysicalDistribute[DistributionSpecGather]
@@ -28,6 +80,7 @@ PhysicalResultSink
----------------------------------------------PhysicalProject
------------------------------------------------filter((date_dim.d_moy = 1)
and (date_dim.d_year = 1999))
--------------------------------------------------PhysicalOlapScan[date_dim]
+<<<<<<< HEAD
--------------------------------PhysicalDistribute[DistributionSpecReplicated]
----------------------------------PhysicalAssertNumRows
------------------------------------PhysicalDistribute[DistributionSpecGather]
@@ -80,4 +133,6 @@ PhysicalResultSink
----------------------------------------PhysicalDistribute[DistributionSpecReplicated]
------------------------------------------PhysicalProject
--------------------------------------------PhysicalOlapScan[store]
+=======
+>>>>>>> 9c7d4b518e ([fix](Nereids): fix AssertNumRows StatsCalculator (#30053))
diff --git
a/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query6.out
b/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query6.out
index b09c0da2bb3..889a34c0eb0 100644
--- a/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query6.out
+++ b/regression-test/data/nereids_tpcds_shape_sf1000_p0/shape/query6.out
@@ -10,44 +10,44 @@ PhysicalResultSink
--------------PhysicalDistribute[DistributionSpecHash]
----------------hashAgg[LOCAL]
------------------PhysicalProject
---------------------hashJoin[INNER_JOIN] hashCondition=((c.c_customer_sk =
s.ss_customer_sk)) otherCondition=() build RFs:RF5
c_customer_sk->[ss_customer_sk]
+--------------------hashJoin[INNER_JOIN] hashCondition=((a.ca_address_sk =
c.c_current_addr_sk)) otherCondition=() build RFs:RF5
c_current_addr_sk->[ca_address_sk]
+----------------------PhysicalProject
+------------------------PhysicalOlapScan[customer_address] apply RFs: RF5
----------------------PhysicalDistribute[DistributionSpecHash]
------------------------PhysicalProject
---------------------------hashJoin[INNER_JOIN] hashCondition=((s.ss_item_sk =
i.i_item_sk)) otherCondition=() build RFs:RF4 i_item_sk->[ss_item_sk]
-----------------------------PhysicalProject
-------------------------------hashJoin[INNER_JOIN]
hashCondition=((s.ss_sold_date_sk = d.d_date_sk)) otherCondition=() build
RFs:RF3 d_date_sk->[ss_sold_date_sk]
---------------------------------PhysicalProject
-----------------------------------PhysicalOlapScan[store_sales] apply RFs: RF3
RF4 RF5
---------------------------------PhysicalDistribute[DistributionSpecReplicated]
-----------------------------------hashJoin[INNER_JOIN]
hashCondition=((d.d_month_seq = date_dim.d_month_seq)) otherCondition=() build
RFs:RF2 d_month_seq->[d_month_seq]
-------------------------------------PhysicalProject
---------------------------------------PhysicalOlapScan[date_dim] apply RFs: RF2
-------------------------------------PhysicalDistribute[DistributionSpecReplicated]
---------------------------------------PhysicalAssertNumRows
-----------------------------------------PhysicalDistribute[DistributionSpecGather]
-------------------------------------------hashAgg[GLOBAL]
---------------------------------------------PhysicalDistribute[DistributionSpecHash]
-----------------------------------------------hashAgg[LOCAL]
-------------------------------------------------PhysicalProject
---------------------------------------------------filter((date_dim.d_moy = 3)
and (date_dim.d_year = 2002))
-----------------------------------------------------PhysicalOlapScan[date_dim]
-----------------------------PhysicalDistribute[DistributionSpecReplicated]
-------------------------------hashJoin[INNER_JOIN]
hashCondition=((j.i_category = i.i_category))
otherCondition=((cast(i_current_price as DECIMALV3(38, 5)) > (1.2 *
avg(cast(i_current_price as DECIMALV3(9, 4)))))) build RFs:RF1
i_category->[i_category]
---------------------------------PhysicalProject
-----------------------------------PhysicalOlapScan[item] apply RFs: RF1
---------------------------------PhysicalDistribute[DistributionSpecReplicated]
-----------------------------------hashAgg[GLOBAL]
-------------------------------------PhysicalDistribute[DistributionSpecHash]
---------------------------------------hashAgg[LOCAL]
-----------------------------------------PhysicalProject
-------------------------------------------PhysicalOlapScan[item]
-----------------------PhysicalDistribute[DistributionSpecHash]
-------------------------PhysicalProject
---------------------------hashJoin[INNER_JOIN] hashCondition=((a.ca_address_sk
= c.c_current_addr_sk)) otherCondition=() build RFs:RF0
ca_address_sk->[c_current_addr_sk]
+--------------------------hashJoin[INNER_JOIN] hashCondition=((c.c_customer_sk
= s.ss_customer_sk)) otherCondition=() build RFs:RF4
ss_customer_sk->[c_customer_sk]
----------------------------PhysicalDistribute[DistributionSpecHash]
------------------------------PhysicalProject
---------------------------------PhysicalOlapScan[customer] apply RFs: RF0
+--------------------------------PhysicalOlapScan[customer] apply RFs: RF4
----------------------------PhysicalDistribute[DistributionSpecHash]
------------------------------PhysicalProject
---------------------------------PhysicalOlapScan[customer_address]
+--------------------------------hashJoin[INNER_JOIN]
hashCondition=((s.ss_item_sk = i.i_item_sk)) otherCondition=() build RFs:RF3
i_item_sk->[ss_item_sk]
+----------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------PhysicalProject
+--------------------------------------hashJoin[INNER_JOIN]
hashCondition=((s.ss_sold_date_sk = d.d_date_sk)) otherCondition=() build
RFs:RF2 d_date_sk->[ss_sold_date_sk]
+----------------------------------------PhysicalProject
+------------------------------------------PhysicalOlapScan[store_sales] apply
RFs: RF2 RF3
+----------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((d.d_month_seq = date_dim.d_month_seq)) otherCondition=() build
RFs:RF1 d_month_seq->[d_month_seq]
+--------------------------------------------PhysicalProject
+----------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF1
+--------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+----------------------------------------------PhysicalAssertNumRows
+------------------------------------------------PhysicalDistribute[DistributionSpecGather]
+--------------------------------------------------hashAgg[GLOBAL]
+----------------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------------------hashAgg[LOCAL]
+--------------------------------------------------------PhysicalProject
+----------------------------------------------------------filter((date_dim.d_moy
= 3) and (date_dim.d_year = 2002))
+------------------------------------------------------------PhysicalOlapScan[date_dim]
+----------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------hashJoin[INNER_JOIN]
hashCondition=((j.i_category = i.i_category))
otherCondition=((cast(i_current_price as DECIMALV3(38, 5)) > (1.2 *
avg(cast(i_current_price as DECIMALV3(9, 4)))))) build RFs:RF0
i_category->[i_category]
+--------------------------------------PhysicalProject
+----------------------------------------PhysicalOlapScan[item] apply RFs: RF0
+--------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+----------------------------------------hashAgg[GLOBAL]
+------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------hashAgg[LOCAL]
+----------------------------------------------PhysicalProject
+------------------------------------------------PhysicalOlapScan[item]
diff --git
a/regression-test/data/nereids_tpcds_shape_sf100_p0/noStatsRfPrune/query58.out
b/regression-test/data/nereids_tpcds_shape_sf100_p0/noStatsRfPrune/query58.out
index c59907df29c..c0082e8cea0 100644
---
a/regression-test/data/nereids_tpcds_shape_sf100_p0/noStatsRfPrune/query58.out
+++
b/regression-test/data/nereids_tpcds_shape_sf100_p0/noStatsRfPrune/query58.out
@@ -28,10 +28,9 @@ PhysicalResultSink
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------PhysicalProject
----------------------------------------hashJoin[INNER_JOIN]
hashCondition=((date_dim.d_week_seq = date_dim.d_week_seq)) otherCondition=()
build RFs:RF8 d_week_seq->[d_week_seq]
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
---------------------------------------------PhysicalProject
-----------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF8
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------PhysicalProject
+--------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF8
+------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
--------------------------------------------PhysicalAssertNumRows
----------------------------------------------PhysicalDistribute[DistributionSpecGather]
------------------------------------------------PhysicalProject
@@ -59,10 +58,9 @@ PhysicalResultSink
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------PhysicalProject
----------------------------------------hashJoin[INNER_JOIN]
hashCondition=((date_dim.d_week_seq = date_dim.d_week_seq)) otherCondition=()
build RFs:RF4 d_week_seq->[d_week_seq]
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
---------------------------------------------PhysicalProject
-----------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF4
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------PhysicalProject
+--------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF4
+------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
--------------------------------------------PhysicalAssertNumRows
----------------------------------------------PhysicalDistribute[DistributionSpecGather]
------------------------------------------------PhysicalProject
@@ -90,10 +88,9 @@ PhysicalResultSink
--------------------------------PhysicalDistribute[DistributionSpecHash]
----------------------------------PhysicalProject
------------------------------------hashJoin[INNER_JOIN]
hashCondition=((date_dim.d_week_seq = date_dim.d_week_seq)) otherCondition=()
build RFs:RF0 d_week_seq->[d_week_seq]
---------------------------------------PhysicalDistribute[DistributionSpecHash]
-----------------------------------------PhysicalProject
-------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF0
---------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------PhysicalProject
+----------------------------------------PhysicalOlapScan[date_dim] apply RFs:
RF0
+--------------------------------------PhysicalDistribute[DistributionSpecReplicated]
----------------------------------------PhysicalAssertNumRows
------------------------------------------PhysicalDistribute[DistributionSpecGather]
--------------------------------------------PhysicalProject
diff --git
a/regression-test/data/nereids_tpcds_shape_sf100_p0/no_stats_shape/query58.out
b/regression-test/data/nereids_tpcds_shape_sf100_p0/no_stats_shape/query58.out
index a449bf2addb..2814b83c308 100644
---
a/regression-test/data/nereids_tpcds_shape_sf100_p0/no_stats_shape/query58.out
+++
b/regression-test/data/nereids_tpcds_shape_sf100_p0/no_stats_shape/query58.out
@@ -28,10 +28,9 @@ PhysicalResultSink
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------PhysicalProject
----------------------------------------hashJoin[INNER_JOIN]
hashCondition=((date_dim.d_week_seq = date_dim.d_week_seq)) otherCondition=()
build RFs:RF8 d_week_seq->[d_week_seq]
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
---------------------------------------------PhysicalProject
-----------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF8 RF9
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------PhysicalProject
+--------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF8 RF9
+------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
--------------------------------------------PhysicalAssertNumRows
----------------------------------------------PhysicalDistribute[DistributionSpecGather]
------------------------------------------------PhysicalProject
@@ -59,10 +58,9 @@ PhysicalResultSink
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------PhysicalProject
----------------------------------------hashJoin[INNER_JOIN]
hashCondition=((date_dim.d_week_seq = date_dim.d_week_seq)) otherCondition=()
build RFs:RF4 d_week_seq->[d_week_seq]
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
---------------------------------------------PhysicalProject
-----------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF4 RF5
-------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------PhysicalProject
+--------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF4 RF5
+------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
--------------------------------------------PhysicalAssertNumRows
----------------------------------------------PhysicalDistribute[DistributionSpecGather]
------------------------------------------------PhysicalProject
@@ -90,10 +88,9 @@ PhysicalResultSink
--------------------------------PhysicalDistribute[DistributionSpecHash]
----------------------------------PhysicalProject
------------------------------------hashJoin[INNER_JOIN]
hashCondition=((date_dim.d_week_seq = date_dim.d_week_seq)) otherCondition=()
build RFs:RF0 d_week_seq->[d_week_seq]
---------------------------------------PhysicalDistribute[DistributionSpecHash]
-----------------------------------------PhysicalProject
-------------------------------------------PhysicalOlapScan[date_dim] apply
RFs: RF0 RF1
---------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------PhysicalProject
+----------------------------------------PhysicalOlapScan[date_dim] apply RFs:
RF0 RF1
+--------------------------------------PhysicalDistribute[DistributionSpecReplicated]
----------------------------------------PhysicalAssertNumRows
------------------------------------------PhysicalDistribute[DistributionSpecGather]
--------------------------------------------PhysicalProject
diff --git
a/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query54.out
b/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query54.out
index 1bacc9f548a..317cf663ecd 100644
--- a/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query54.out
+++ b/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query54.out
@@ -13,10 +13,11 @@ PhysicalResultSink
--------------------PhysicalDistribute[DistributionSpecHash]
----------------------hashAgg[LOCAL]
------------------------PhysicalProject
---------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) <= (d_month_seq + 3))
+--------------------------hashJoin[INNER_JOIN]
hashCondition=((store_sales.ss_sold_date_sk = date_dim.d_date_sk))
otherCondition=() build RFs:RF7 d_date_sk->[ss_sold_date_sk]
----------------------------PhysicalProject
-------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) >= (d_month_seq + 1))
+------------------------------hashJoin[INNER_JOIN]
hashCondition=((my_customers.c_customer_sk = store_sales.ss_customer_sk))
otherCondition=() build RFs:RF6 c_customer_sk->[ss_customer_sk]
--------------------------------PhysicalProject
+<<<<<<< HEAD
----------------------------------hashJoin[INNER_JOIN]
hashCondition=((store_sales.ss_sold_date_sk = date_dim.d_date_sk))
otherCondition=()
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------PhysicalProject
@@ -79,7 +80,69 @@ PhysicalResultSink
----------------------------------hashAgg[GLOBAL]
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------hashAgg[LOCAL]
+=======
+----------------------------------PhysicalOlapScan[store_sales] apply RFs: RF6
RF7
+--------------------------------PhysicalDistribute[DistributionSpecReplicated]
+----------------------------------PhysicalProject
+------------------------------------hashJoin[INNER_JOIN]
hashCondition=((customer_address.ca_county = store.s_county) and
(customer_address.ca_state = store.s_state)) otherCondition=() build RFs:RF4
s_county->[ca_county];RF5 s_state->[ca_state]
+--------------------------------------hashJoin[INNER_JOIN]
hashCondition=((my_customers.c_current_addr_sk =
customer_address.ca_address_sk)) otherCondition=() build RFs:RF3
c_current_addr_sk->[ca_address_sk]
----------------------------------------PhysicalProject
-------------------------------------------filter((date_dim.d_moy = 5) and
(date_dim.d_year = 1998))
---------------------------------------------PhysicalOlapScan[date_dim]
+------------------------------------------PhysicalOlapScan[customer_address]
apply RFs: RF3 RF4 RF5
+----------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------PhysicalProject
+--------------------------------------------hashAgg[GLOBAL]
+----------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------------hashAgg[LOCAL]
+--------------------------------------------------PhysicalProject
+----------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((customer.c_customer_sk = cs_or_ws_sales.customer_sk))
otherCondition=() build RFs:RF2 customer_sk->[c_customer_sk]
+------------------------------------------------------PhysicalProject
+--------------------------------------------------------PhysicalOlapScan[customer]
apply RFs: RF2
+------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------------------------PhysicalProject
+----------------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((cs_or_ws_sales.sold_date_sk = date_dim.d_date_sk))
otherCondition=() build RFs:RF1 d_date_sk->[cs_sold_date_sk,ws_sold_date_sk]
+------------------------------------------------------------PhysicalProject
+--------------------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((cs_or_ws_sales.item_sk = item.i_item_sk)) otherCondition=()
build RFs:RF0 i_item_sk->[cs_item_sk,ws_item_sk]
+----------------------------------------------------------------PhysicalUnion
+------------------------------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------------PhysicalOlapScan[catalog_sales]
apply RFs: RF0 RF1
+------------------------------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------------PhysicalOlapScan[web_sales]
apply RFs: RF0 RF1
+----------------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------------------------------------------PhysicalProject
+--------------------------------------------------------------------filter((item.i_category
= 'Women') and (item.i_class = 'maternity'))
+----------------------------------------------------------------------PhysicalOlapScan[item]
+------------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------filter((date_dim.d_moy
= 5) and (date_dim.d_year = 1998))
+------------------------------------------------------------------PhysicalOlapScan[date_dim]
+--------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+>>>>>>> 9c7d4b518e ([fix](Nereids): fix AssertNumRows StatsCalculator (#30053))
+----------------------------------------PhysicalProject
+------------------------------------------PhysicalOlapScan[store]
+----------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------PhysicalProject
+--------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) <= (d_month_seq + 3))
+----------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq
as BIGINT) >= (d_month_seq + 1))
+------------------------------------PhysicalProject
+--------------------------------------PhysicalOlapScan[date_dim]
+------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------PhysicalAssertNumRows
+----------------------------------------PhysicalDistribute[DistributionSpecGather]
+------------------------------------------hashAgg[GLOBAL]
+--------------------------------------------PhysicalDistribute[DistributionSpecHash]
+----------------------------------------------hashAgg[LOCAL]
+------------------------------------------------PhysicalProject
+--------------------------------------------------filter((date_dim.d_moy = 5)
and (date_dim.d_year = 1998))
+----------------------------------------------------PhysicalOlapScan[date_dim]
+----------------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------------PhysicalAssertNumRows
+--------------------------------------PhysicalDistribute[DistributionSpecGather]
+----------------------------------------hashAgg[GLOBAL]
+------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------hashAgg[LOCAL]
+----------------------------------------------PhysicalProject
+------------------------------------------------filter((date_dim.d_moy = 5)
and (date_dim.d_year = 1998))
+--------------------------------------------------PhysicalOlapScan[date_dim]
diff --git
a/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query6.out
b/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query6.out
index 07fa54e46ba..28f53e05bbe 100644
--- a/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query6.out
+++ b/regression-test/data/nereids_tpcds_shape_sf100_p0/rf_prune/query6.out
@@ -10,7 +10,9 @@ PhysicalResultSink
--------------PhysicalDistribute[DistributionSpecHash]
----------------hashAgg[LOCAL]
------------------PhysicalProject
---------------------hashJoin[INNER_JOIN] hashCondition=((a.ca_address_sk =
c.c_current_addr_sk)) otherCondition=()
+--------------------hashJoin[INNER_JOIN] hashCondition=((a.ca_address_sk =
c.c_current_addr_sk)) otherCondition=() build RFs:RF5
c_current_addr_sk->[ca_address_sk]
+----------------------PhysicalProject
+------------------------PhysicalOlapScan[customer_address] apply RFs: RF5
----------------------PhysicalDistribute[DistributionSpecHash]
------------------------PhysicalProject
--------------------------hashJoin[INNER_JOIN] hashCondition=((c.c_customer_sk
= s.ss_customer_sk)) otherCondition=() build RFs:RF4
ss_customer_sk->[c_customer_sk]
@@ -48,7 +50,4 @@ PhysicalResultSink
--------------------------------------------hashAgg[LOCAL]
----------------------------------------------PhysicalProject
------------------------------------------------PhysicalOlapScan[item]
-----------------------PhysicalDistribute[DistributionSpecHash]
-------------------------PhysicalProject
---------------------------PhysicalOlapScan[customer_address]
diff --git
a/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query54.out
b/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query54.out
index b4f27acdc19..258f66e9c6a 100644
--- a/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query54.out
+++ b/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query54.out
@@ -13,10 +13,11 @@ PhysicalResultSink
--------------------PhysicalDistribute[DistributionSpecHash]
----------------------hashAgg[LOCAL]
------------------------PhysicalProject
---------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) <= (d_month_seq + 3))
+--------------------------hashJoin[INNER_JOIN]
hashCondition=((store_sales.ss_sold_date_sk = date_dim.d_date_sk))
otherCondition=() build RFs:RF7 d_date_sk->[ss_sold_date_sk]
----------------------------PhysicalProject
-------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) >= (d_month_seq + 1))
+------------------------------hashJoin[INNER_JOIN]
hashCondition=((my_customers.c_customer_sk = store_sales.ss_customer_sk))
otherCondition=() build RFs:RF6 c_customer_sk->[ss_customer_sk]
--------------------------------PhysicalProject
+<<<<<<< HEAD
----------------------------------hashJoin[INNER_JOIN]
hashCondition=((store_sales.ss_sold_date_sk = date_dim.d_date_sk))
otherCondition=() build RFs:RF6 d_date_sk->[ss_sold_date_sk]
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------PhysicalProject
@@ -79,7 +80,69 @@ PhysicalResultSink
----------------------------------hashAgg[GLOBAL]
------------------------------------PhysicalDistribute[DistributionSpecHash]
--------------------------------------hashAgg[LOCAL]
+=======
+----------------------------------PhysicalOlapScan[store_sales] apply RFs: RF6
RF7
+--------------------------------PhysicalDistribute[DistributionSpecReplicated]
+----------------------------------PhysicalProject
+------------------------------------hashJoin[INNER_JOIN]
hashCondition=((customer_address.ca_county = store.s_county) and
(customer_address.ca_state = store.s_state)) otherCondition=() build RFs:RF4
s_county->[ca_county];RF5 s_state->[ca_state]
+--------------------------------------hashJoin[INNER_JOIN]
hashCondition=((my_customers.c_current_addr_sk =
customer_address.ca_address_sk)) otherCondition=() build RFs:RF3
c_current_addr_sk->[ca_address_sk]
----------------------------------------PhysicalProject
-------------------------------------------filter((date_dim.d_moy = 5) and
(date_dim.d_year = 1998))
---------------------------------------------PhysicalOlapScan[date_dim]
+------------------------------------------PhysicalOlapScan[customer_address]
apply RFs: RF3 RF4 RF5
+----------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------PhysicalProject
+--------------------------------------------hashAgg[GLOBAL]
+----------------------------------------------PhysicalDistribute[DistributionSpecHash]
+------------------------------------------------hashAgg[LOCAL]
+--------------------------------------------------PhysicalProject
+----------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((customer.c_customer_sk = cs_or_ws_sales.customer_sk))
otherCondition=() build RFs:RF2 customer_sk->[c_customer_sk]
+------------------------------------------------------PhysicalProject
+--------------------------------------------------------PhysicalOlapScan[customer]
apply RFs: RF2
+------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------------------------PhysicalProject
+----------------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((cs_or_ws_sales.sold_date_sk = date_dim.d_date_sk))
otherCondition=() build RFs:RF1 d_date_sk->[cs_sold_date_sk,ws_sold_date_sk]
+------------------------------------------------------------PhysicalProject
+--------------------------------------------------------------hashJoin[INNER_JOIN]
hashCondition=((cs_or_ws_sales.item_sk = item.i_item_sk)) otherCondition=()
build RFs:RF0 i_item_sk->[cs_item_sk,ws_item_sk]
+----------------------------------------------------------------PhysicalUnion
+------------------------------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------------PhysicalOlapScan[catalog_sales]
apply RFs: RF0 RF1
+------------------------------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------------PhysicalOlapScan[web_sales]
apply RFs: RF0 RF1
+----------------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------------------------------------------PhysicalProject
+--------------------------------------------------------------------filter((item.i_category
= 'Women') and (item.i_class = 'maternity'))
+----------------------------------------------------------------------PhysicalOlapScan[item]
+------------------------------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------------------------------PhysicalProject
+----------------------------------------------------------------filter((date_dim.d_moy
= 5) and (date_dim.d_year = 1998))
+------------------------------------------------------------------PhysicalOlapScan[date_dim]
+--------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+>>>>>>> 9c7d4b518e ([fix](Nereids): fix AssertNumRows StatsCalculator (#30053))
+----------------------------------------PhysicalProject
+------------------------------------------PhysicalOlapScan[store]
+----------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------PhysicalProject
+--------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq as
BIGINT) <= (d_month_seq + 3))
+----------------------------------NestedLoopJoin[INNER_JOIN](cast(d_month_seq
as BIGINT) >= (d_month_seq + 1))
+------------------------------------PhysicalProject
+--------------------------------------PhysicalOlapScan[date_dim]
+------------------------------------PhysicalDistribute[DistributionSpecReplicated]
+--------------------------------------PhysicalAssertNumRows
+----------------------------------------PhysicalDistribute[DistributionSpecGather]
+------------------------------------------hashAgg[GLOBAL]
+--------------------------------------------PhysicalDistribute[DistributionSpecHash]
+----------------------------------------------hashAgg[LOCAL]
+------------------------------------------------PhysicalProject
+--------------------------------------------------filter((date_dim.d_moy = 5)
and (date_dim.d_year = 1998))
+----------------------------------------------------PhysicalOlapScan[date_dim]
+----------------------------------PhysicalDistribute[DistributionSpecReplicated]
+------------------------------------PhysicalAssertNumRows
+--------------------------------------PhysicalDistribute[DistributionSpecGather]
+----------------------------------------hashAgg[GLOBAL]
+------------------------------------------PhysicalDistribute[DistributionSpecHash]
+--------------------------------------------hashAgg[LOCAL]
+----------------------------------------------PhysicalProject
+------------------------------------------------filter((date_dim.d_moy = 5)
and (date_dim.d_year = 1998))
+--------------------------------------------------PhysicalOlapScan[date_dim]
diff --git a/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query6.out
b/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query6.out
index 000b390d023..889a34c0eb0 100644
--- a/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query6.out
+++ b/regression-test/data/nereids_tpcds_shape_sf100_p0/shape/query6.out
@@ -10,13 +10,15 @@ PhysicalResultSink
--------------PhysicalDistribute[DistributionSpecHash]
----------------hashAgg[LOCAL]
------------------PhysicalProject
---------------------hashJoin[INNER_JOIN] hashCondition=((a.ca_address_sk =
c.c_current_addr_sk)) otherCondition=() build RFs:RF5
ca_address_sk->[c_current_addr_sk]
+--------------------hashJoin[INNER_JOIN] hashCondition=((a.ca_address_sk =
c.c_current_addr_sk)) otherCondition=() build RFs:RF5
c_current_addr_sk->[ca_address_sk]
+----------------------PhysicalProject
+------------------------PhysicalOlapScan[customer_address] apply RFs: RF5
----------------------PhysicalDistribute[DistributionSpecHash]
------------------------PhysicalProject
--------------------------hashJoin[INNER_JOIN] hashCondition=((c.c_customer_sk
= s.ss_customer_sk)) otherCondition=() build RFs:RF4
ss_customer_sk->[c_customer_sk]
----------------------------PhysicalDistribute[DistributionSpecHash]
------------------------------PhysicalProject
---------------------------------PhysicalOlapScan[customer] apply RFs: RF4 RF5
+--------------------------------PhysicalOlapScan[customer] apply RFs: RF4
----------------------------PhysicalDistribute[DistributionSpecHash]
------------------------------PhysicalProject
--------------------------------hashJoin[INNER_JOIN]
hashCondition=((s.ss_item_sk = i.i_item_sk)) otherCondition=() build RFs:RF3
i_item_sk->[ss_item_sk]
@@ -48,7 +50,4 @@ PhysicalResultSink
--------------------------------------------hashAgg[LOCAL]
----------------------------------------------PhysicalProject
------------------------------------------------PhysicalOlapScan[item]
-----------------------PhysicalDistribute[DistributionSpecHash]
-------------------------PhysicalProject
---------------------------PhysicalOlapScan[customer_address]
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
