[test](migrate) move some cases from p2 to p0 (#36750)(#36787)

.licenserc.yaml

@@ -82,6 +82,7 @@ header:
     - "docker/thirdparties/docker-compose/hive/scripts/create_tpch1_orc.hql"
     - "docker/thirdparties/docker-compose/hive/scripts/create_tpch1_parquet.hql"
     - "docker/thirdparties/docker-compose/hive/scripts/preinstalled_data/"
+    - "docker/thirdparties/docker-compose/hive/scripts/data/**"
     - "docker/thirdparties/docker-compose/iceberg/spark-defaults.conf.tpl"
     - "conf/mysql_ssl_default_certificate/*"
     - "conf/mysql_ssl_default_certificate/client_certificate/ca.pem"

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/datev2_csv/create_table.hql

@@ -0,0 +1,20 @@
+create database if not exists multi_catalog;
+
+use multi_catalog;
+
+CREATE external TABLE `datev2_csv`(
+  `id` int,
+  `day` date)
+ROW FORMAT SERDE
+  'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.mapred.TextInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
+LOCATION
+  '/user/doris/suites/multi_catalog/datev2_csv'
+TBLPROPERTIES (
+  'transient_lastDdlTime'='1688118691');
+
+msck repair table datev2_csv;
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/datev2_csv/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -x
+
+CUR_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
+
+## mkdir and put data to hdfs
+cd "${CUR_DIR}" && rm -rf data/ && tar xzf data.tar.gz
+hadoop fs -mkdir -p /user/doris/suites/multi_catalog/
+hadoop fs -put "${CUR_DIR}"/data/* /user/doris/suites/multi_catalog/
+
+# create table
+hive -f "${CUR_DIR}/create_table.hql"
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/datev2_orc/create_table.hql

@@ -0,0 +1,20 @@
+create database if not exists multi_catalog;
+
+use multi_catalog;
+
+CREATE external TABLE `datev2_orc`(
+  `id` int,
+  `day` date)
+ROW FORMAT SERDE
+  'org.apache.hadoop.hive.ql.io.orc.OrcSerde'
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.orc.OrcInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat'
+LOCATION
+  '/user/doris/suites/multi_catalog/datev2_orc'
+TBLPROPERTIES (
+  'transient_lastDdlTime'='1688118707');
+
+msck repair table datev2_orc;
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/datev2_orc/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -x
+
+CUR_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
+
+## mkdir and put data to hdfs
+cd "${CUR_DIR}" && rm -rf data/ && tar xzf data.tar.gz
+hadoop fs -mkdir -p /user/doris/suites/multi_catalog/
+hadoop fs -put "${CUR_DIR}"/data/* /user/doris/suites/multi_catalog/
+
+# create table
+hive -f "${CUR_DIR}"/create_table.hql
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/datev2_parquet/create_table.hql

@@ -0,0 +1,20 @@
+create database if not exists multi_catalog;
+
+use multi_catalog;
+
+CREATE external TABLE `datev2_parquet`(
+  `id` int,
+  `day` date)
+ROW FORMAT SERDE
+  'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
+LOCATION
+  '/user/doris/suites/multi_catalog/datev2_parquet'
+TBLPROPERTIES (
+  'transient_lastDdlTime'='1688118725');
+
+msck repair table datev2_parquet;
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/datev2_parquet/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -x
+
+CUR_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
+
+## mkdir and put data to hdfs
+cd "${CUR_DIR}" && rm -rf data/ && tar xzf data.tar.gz
+hadoop fs -mkdir -p /user/doris/suites/multi_catalog/
+hadoop fs -put "${CUR_DIR}"/data/* /user/doris/suites/multi_catalog/
+
+# create table
+hive -f "${CUR_DIR}"/create_table.hql
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/hive_upper_case_orc/create_table.hql

@@ -0,0 +1,21 @@
+create database if not exists multi_catalog;
+use multi_catalog;
+
+CREATE TABLE `hive_upper_case_orc`(
+  `id` int,
+  `name` string)
+ROW FORMAT SERDE
+  'org.apache.hadoop.hive.ql.io.orc.OrcSerde'
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.orc.OrcInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat'
+LOCATION
+  '/user/doris/suites/multi_catalog/hive_upper_case_orc'
+TBLPROPERTIES (
+  'spark.sql.create.version'='3.2.1',
+  'spark.sql.sources.schema'='{"type":"struct","fields":[{"name":"ID","type":"integer","nullable":true,"metadata":{}},{"name":"NAME","type":"string","nullable":true,"metadata":{}}]}',
+  'transient_lastDdlTime'='1674189057');
+
+msck repair table hive_upper_case_orc;
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/hive_upper_case_orc/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -x
+
+CUR_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
+
+## mkdir and put data to hdfs
+cd "${CUR_DIR}" && rm -rf data/ && tar xzf data.tar.gz
+hadoop fs -mkdir -p /user/doris/suites/multi_catalog/
+hadoop fs -put "${CUR_DIR}"/data/* /user/doris/suites/multi_catalog/
+
+# create table
+hive -f "${CUR_DIR}"/create_table.hql
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/hive_upper_case_parquet/create_table.hql

@@ -0,0 +1,21 @@
+create database if not exists multi_catalog;
+use multi_catalog;
+
+CREATE TABLE `hive_upper_case_parquet`(
+  `id` int,
+  `name` string)
+ROW FORMAT SERDE
+  'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
+LOCATION
+  '/user/doris/suites/multi_catalog/hive_upper_case_parquet'
+TBLPROPERTIES (
+  'spark.sql.create.version'='3.2.1',
+  'spark.sql.sources.schema'='{"type":"struct","fields":[{"name":"ID","type":"integer","nullable":true,"metadata":{}},{"name":"NAME","type":"string","nullable":true,"metadata":{}}]}',
+  'transient_lastDdlTime'='1674189051');
+
+msck repair table hive_upper_case_parquet;
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/hive_upper_case_parquet/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -x
+
+CUR_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
+
+## mkdir and put data to hdfs
+cd "${CUR_DIR}" && rm -rf data/ && tar xzf data.tar.gz
+hadoop fs -mkdir -p /user/doris/suites/multi_catalog/
+hadoop fs -put "${CUR_DIR}"/data/* /user/doris/suites/multi_catalog/
+
+# create table
+hive -f "${CUR_DIR}"/create_table.hql
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/parquet_lz4_compression/create_table.hql

@@ -0,0 +1,31 @@
+create database if not exists multi_catalog;
+use multi_catalog;
+
+CREATE TABLE `parquet_lz4_compression`(
+  `col_int` int,
+  `col_smallint` smallint,
+  `col_tinyint` tinyint,
+  `col_bigint` bigint,
+  `col_float` float,
+  `col_double` double,
+  `col_boolean` boolean,
+  `col_string` string,
+  `col_char` char(10),
+  `col_varchar` varchar(25),
+  `col_date` date,
+  `col_timestamp` timestamp,
+  `col_decimal` decimal(10,2))
+ROW FORMAT SERDE
+  'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
+LOCATION
+  '/user/doris/suites/multi_catalog/parquet_lz4_compression'
+TBLPROPERTIES (
+  'parquet.compression'='LZ4',
+  'transient_lastDdlTime'='1700723950');
+
+msck repair table parquet_lz4_compression;
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/parquet_lz4_compression/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -x
+
+CUR_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
+
+## mkdir and put data to hdfs
+cd "${CUR_DIR}" && rm -rf data/ && tar xzf data.tar.gz
+hadoop fs -mkdir -p /user/doris/suites/multi_catalog/
+hadoop fs -put "${CUR_DIR}"/data/* /user/doris/suites/multi_catalog/
+
+# create table
+hive -f "${CUR_DIR}"/create_table.hql
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/parquet_lzo_compression/create_table.hql

@@ -0,0 +1,31 @@
+create database if not exists multi_catalog;
+use multi_catalog;
+
+CREATE TABLE `parquet_lzo_compression`(
+  `col_int` int,
+  `col_smallint` smallint,
+  `col_tinyint` tinyint,
+  `col_bigint` bigint,
+  `col_float` float,
+  `col_double` double,
+  `col_boolean` boolean,
+  `col_string` string,
+  `col_char` char(10),
+  `col_varchar` varchar(25),
+  `col_date` date,
+  `col_timestamp` timestamp,
+  `col_decimal` decimal(10,2))
+ROW FORMAT SERDE
+  'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
+LOCATION
+  '/user/doris/suites/multi_catalog/parquet_lzo_compression'
+TBLPROPERTIES (
+  'parquet.compression'='LZO',
+  'transient_lastDdlTime'='1701173147');
+
+msck repair table parquet_lzo_compression;
+

docker/thirdparties/docker-compose/hive/scripts/data/multi_catalog/parquet_lzo_compression/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -x
+
+CUR_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &>/dev/null && pwd)"
+
+## mkdir and put data to hdfs
+cd "${CUR_DIR}" && rm -rf data/ && tar xzf data.tar.gz
+hadoop fs -mkdir -p /user/doris/suites/multi_catalog/
+hadoop fs -put "${CUR_DIR}"/data/* /user/doris/suites/multi_catalog/
+
+# create table
+hive -f "${CUR_DIR}"/create_table.hql
+