HIVE-29361: BasicStatsTask.aggregateStats shouldn't update partition stats in HMS for non-native tables

iceberg/iceberg-handler/src/test/queries/positive/col_stats.q

@@ -3,58 +3,105 @@
 set hive.stats.autogather=true;
 set hive.stats.column.autogather=true;
 
+-- Create source table
+drop table if exists src_ice;
+create external table src_ice(
+    a int,
+    b string,
+    c int)
+stored by iceberg;
+
+insert into src_ice values
+    (1, 'one', 50),
+    (2, 'two', 51),
+    (2, 'two', 51),
+    (2, 'two', 51),
+    (3, 'three', 52),
+    (4, 'four', 53),
+    (5, 'five', 54),
+    (111, 'one', 55),
+    (333, 'two', 56);
+
+-- Test hive.iceberg.stats.source = iceberg
 set hive.iceberg.stats.source=iceberg;
+
+
+-- Test NON-PARTITIONED table with hive.iceberg.stats.source=iceberg
 drop table if exists tbl_ice_puffin;
-create external table tbl_ice_puffin(a int, b string, c int) stored by iceberg tblproperties ('format-version'='2');
-insert into tbl_ice_puffin values (1, 'one', 50), (2, 'two', 51),(2, 'two', 51),(2, 'two', 51), (3, 'three', 52), (4, 'four', 53), (5, 'five', 54), (111, 'one', 55), (333, 'two', 56);
-insert into tbl_ice_puffin values (1, 'one', 50), (2, 'two', 51),(2, 'two', 51),(2, 'two', 51), (3, 'three', 52), (4, 'four', 53), (5, 'five', 54), (111, 'one', 55), (333, 'two', 56);
-explain select * from tbl_ice_puffin order by a, b, c;
-select * from tbl_ice_puffin order by a, b, c;
-desc formatted tbl_ice_puffin b;
-update tbl_ice_puffin set b='two' where b='one' or b='three';
+create external table tbl_ice_puffin(
+    a int,
+    b string,
+    c int)
+stored by iceberg;
+
+insert into tbl_ice_puffin select * from src_ice;
+insert into tbl_ice_puffin select * from src_ice;
+
+select count(*) from tbl_ice_puffin;
+EXPLAIN select min(a), count(distinct b), max(c) from tbl_ice_puffin;
+desc formatted tbl_ice_puffin B;
+
+update tbl_ice_puffin
+    set b='two' where b='one' or b='three';
+
 analyze table tbl_ice_puffin compute statistics for columns;
-explain select * from tbl_ice_puffin order by a, b, c;
-select * from tbl_ice_puffin order by a, b, c;
+
 select count(*) from tbl_ice_puffin;
-desc formatted tbl_ice_puffin b;
+EXPLAIN select min(a), count(distinct b), max(c) from tbl_ice_puffin;
+desc formatted tbl_ice_puffin B;
+
+
+-- Test PARTITIONED table with hive.iceberg.stats.source=iceberg
+drop table tbl_ice_puffin;
+create external table tbl_ice_puffin(
+    a int,
+    b string
+)
+partitioned by (c int)
+stored by iceberg;
 
+insert overwrite table tbl_ice_puffin select * from src_ice;
+delete from tbl_ice_puffin where a <= 2;
 
--- Test if hive.iceberg.stats.source is empty
+analyze table tbl_ice_puffin compute statistics for columns A, C;
+
+select count(*) from tbl_ice_puffin;
+EXPLAIN select min(a), max(c) from tbl_ice_puffin;
+desc formatted tbl_ice_puffin C;
+
+set hive.stats.fetch.column.stats=true;
+EXPLAIN select count(*) from src_ice t1 join tbl_ice_puffin t2 on (t1.a = t2.a);
+
+
+-- Test hive.iceberg.stats.source is empty
 set hive.iceberg.stats.source= ;
-drop table if exists tbl_ice_puffin;
-create external table tbl_ice_puffin(a int, b string, c int) stored by iceberg tblproperties ('format-version'='2');
-insert into tbl_ice_puffin values (1, 'one', 50), (2, 'two', 51),(2, 'two', 51),(2, 'two', 51), (3, 'three', 52), (4, 'four', 53), (5, 'five', 54), (111, 'one', 55), (333, 'two', 56);
-explain select * from tbl_ice_puffin order by a, b, c;
 
+drop table tbl_ice_puffin;
+create external table tbl_ice_puffin(
+    a int,
+    b string,
+    c int)
+stored by iceberg;
+
+insert into tbl_ice_puffin select * from src_ice;
 
-set hive.iceberg.stats.source=iceberg;
-drop table if exists tbl_ice_puffin;
-create external table tbl_ice_puffin(a int, b string, c int) stored by iceberg tblproperties ('format-version'='2');
-insert into tbl_ice_puffin values (1, 'one', 50), (2, 'two', 51),(2, 'two', 51),(2, 'two', 51), (3, 'three', 52), (4, 'four', 53), (5, 'five', 54), (111, 'one', 55), (333, 'two', 56);
-explain select * from tbl_ice_puffin order by a, b, c;
-select * from tbl_ice_puffin order by a, b, c;
 select count(*) from tbl_ice_puffin;
-desc formatted tbl_ice_puffin a;
+EXPLAIN select min(a), count(distinct b), max(c) from tbl_ice_puffin;
+desc formatted tbl_ice_puffin A;
 
 
+-- Test hive.iceberg.stats.source = metastore
 set hive.iceberg.stats.source=metastore;
 
 drop table if exists tbl_ice;
-create external table tbl_ice(a int, b string, c int) stored by iceberg tblproperties ('format-version'='2');
-insert into tbl_ice values (1, 'one', 50), (2, 'two', 51),(2, 'two', 51),(2, 'two', 51), (3, 'three', 52), (4, 'four', 53), (5, 'five', 54), (111, 'one', 55), (333, 'two', 56);
-explain select * from tbl_ice order by a, b, c;
-select * from tbl_ice order by a, b, c;
-select count(*) from tbl_ice;
+create external table tbl_ice(
+    a int,
+    b string,
+    c int)
+stored by iceberg;
 
-set hive.iceberg.stats.source=iceberg;
-delete from tbl_ice_puffin where a = 2;
-analyze table tbl_ice_puffin compute statistics for columns A, C;
-explain select * from tbl_ice_puffin order by a, b, c;
-select count(*) from tbl_ice_puffin;
-desc formatted tbl_ice_puffin C;
+insert into tbl_ice select * from src_ice;
 
-create table t1 (a int) stored by iceberg tblproperties ('format-version'='2');
-create table t2 (b int) stored by iceberg tblproperties ('format-version'='2');
-describe formatted t1;
-describe formatted t2;
-explain select * from t1 join t2 on t1.a = t2.b;
+select count(*) from tbl_ice;
+EXPLAIN select min(a), count(distinct b), max(c) from tbl_ice;
+desc formatted tbl_ice A;

iceberg/iceberg-handler/src/test/queries/positive/puffin_col_stats.q

@@ -1,21 +1,47 @@
 -- Mask random uuid
 --! qt:replace:/(\s+uuid\s+)\S+(\s*)/$1#Masked#$2/
+set hive.fetch.task.conversion=none;
+
 set hive.stats.autogather=true;
 set hive.stats.column.autogather=true;
 
 set hive.iceberg.stats.source=iceberg;
+
 drop table if exists tbl_ice_puffin;
-create external table tbl_ice_puffin(a int, b string, c int) stored by iceberg tblproperties ('format-version'='2');
-insert into tbl_ice_puffin values (1, 'one', 50), (2, 'two', 51),(2, 'two', 51),(2, 'two', 51), (3, 'three', 52), (4, 'four', 53), (5, 'five', 54), (111, 'one', 55), (333, 'two', 56);
+create external table tbl_ice_puffin(
+    a int,
+    b string,
+    c int)
+stored by iceberg;
+
+insert into tbl_ice_puffin values
+    (1, 'one', 50),
+    (2, 'two', 51),
+    (2, 'two', 51),
+    (2, 'two', 51),
+    (3, 'three', 52),
+    (4, 'four', 53),
+    (5, 'five', 54),
+    (111, 'one', 55),
+    (333, 'two', 56);
+
+explain select * from tbl_ice_puffin;
 desc formatted tbl_ice_puffin a;
+desc formatted tbl_ice_puffin b;
 desc formatted tbl_ice_puffin c;
-explain select * from tbl_ice_puffin order by a, b, c;
-insert into tbl_ice_puffin values (1000, 'one', 1000), (5000, 'two', 5000);
+
+insert into tbl_ice_puffin values (1000, 'six', 1000), (5000, 'two', 5000);
+
+explain select * from tbl_ice_puffin;
 desc formatted tbl_ice_puffin a;
+desc formatted tbl_ice_puffin b;
 desc formatted tbl_ice_puffin c;
-explain select * from tbl_ice_puffin order by a, b, c;
-insert into tbl_ice_puffin values (10, 'one', 100000), (5000, 'two', 510000);
-explain select * from tbl_ice_puffin order by a, b, c;
+
+insert into tbl_ice_puffin values (10, 'six', 100000), (5000, 'two', 510000);
+
+explain select * from tbl_ice_puffin;
 desc formatted tbl_ice_puffin a;
+desc formatted tbl_ice_puffin b;
 desc formatted tbl_ice_puffin c;
+
 -- Result:  a = (min: 1, max: 5000) , c =(min: 50, max: 51000)

iceberg/iceberg-handler/src/test/queries/positive/puffin_col_stats_with_time_travel.q

@@ -1,25 +1,33 @@
 set hive.fetch.task.conversion=none;
 
-create external table default.tbl_ice_puffin_time_travel(a int, b string, c int) stored by iceberg;
-insert into default.tbl_ice_puffin_time_travel values (1, 'one', 50), (2, 'two', 51);
-alter table default.tbl_ice_puffin_time_travel create tag checkpoint;
+create external table tbl_ice_puffin_time_travel(
+    a int,
+    b string,
+    c int)
+stored by iceberg;
 
-explain select * from default.tbl_ice_puffin_time_travel;
+insert into tbl_ice_puffin_time_travel values
+    (1, 'one', 50),
+    (2, 'two', 51);
+
+alter table tbl_ice_puffin_time_travel create tag checkpoint;
+
+explain select * from tbl_ice_puffin_time_travel;
 explain select * from default.tbl_ice_puffin_time_travel.tag_checkpoint;
 
 insert into tbl_ice_puffin_time_travel values
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null),
-(null, null, null);
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null),
+    (null, null, null);
 
-explain select * from default.tbl_ice_puffin_time_travel;
+explain select * from tbl_ice_puffin_time_travel;
 explain select * from default.tbl_ice_puffin_time_travel.tag_checkpoint;