Add workload for hive iceberg insertion

Author: PingLiuPing

benchmarks/insert/hive-tpch-parquet/cleanup_sf10.sql

@@ -0,0 +1,9 @@
+USE tpch_sf10_load_iceberg;
+DROP TABLE customer;
+DROP TABLE orders;
+DROP TABLE lineitem;
+DROP TABLE part;
+DROP TABLE supplier;
+DROP TABLE partsupp;
+DROP TABLE nation;
+DROP TABLE region;

benchmarks/insert/hive-tpch-parquet/customer_sf10.sql

@@ -0,0 +1,13 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into customer
+INSERT INTO customer
+SELECT
+    custkey,
+    CAST(name AS VARCHAR(25)),
+    CAST(address AS VARCHAR(40)),
+    nationkey,
+    CAST(phone AS VARCHAR(15)),
+    CAST(acctbal AS DECIMAL(12,2)),
+    CAST(mktsegment AS VARCHAR(10)),
+    CAST(comment AS VARCHAR(117))
+FROM tpch.sf1.customer;

benchmarks/insert/hive-tpch-parquet/hive_ddl_iceberg.sql

@@ -0,0 +1,95 @@
+CREATE SCHEMA IF NOT EXISTS tpch_sf10_load_iceberg WITH (LOCATION = 's3a://presto-workload/tpch-sf10-load-iceberg/');
+USE tpch_sf10_load_iceberg;
+
+-- Creating customer table with PARQUET format
+CREATE TABLE customer (
+    custkey     BIGINT NOT NULL,
+    name        VARCHAR(25) NOT NULL,
+    address     VARCHAR(40) NOT NULL,
+    nationkey   BIGINT NOT NULL,
+    phone       VARCHAR(15) NOT NULL,
+    acctbal     DECIMAL(12,2) NOT NULL,
+    mktsegment  VARCHAR(10) NOT NULL,
+    comment     VARCHAR(117) NOT NULL
+) WITH (format = 'PARQUET');
+
+-- Creating orders table with PARQUET format
+CREATE TABLE orders (
+    orderkey       BIGINT NOT NULL,
+    custkey        BIGINT NOT NULL,
+    orderstatus    VARCHAR(1) NOT NULL,
+    totalprice     DECIMAL(12,2) NOT NULL,
+    orderdate      DATE NOT NULL,
+    orderpriority  VARCHAR(15) NOT NULL,
+    clerk          VARCHAR(15) NOT NULL,
+    shippriority   BIGINT NOT NULL,
+    comment        VARCHAR(79) NOT NULL
+) WITH (format = 'PARQUET');
+
+-- Creating lineitem table with PARQUET format
+CREATE TABLE lineitem (
+    orderkey      BIGINT NOT NULL,
+    partkey       BIGINT NOT NULL,
+    suppkey       BIGINT NOT NULL,
+    linenumber    BIGINT NOT NULL,
+    quantity      DECIMAL(12,2) NOT NULL,
+    extendedprice DECIMAL(12,2) NOT NULL,
+    discount      DECIMAL(12,2) NOT NULL,
+    tax           DECIMAL(12,2) NOT NULL,
+    returnflag    VARCHAR(1) NOT NULL,
+    linestatus    VARCHAR(1) NOT NULL,
+    shipdate      DATE NOT NULL,
+    commitdate    DATE NOT NULL,
+    receiptdate   DATE NOT NULL,
+    shipinstruct  VARCHAR(25) NOT NULL,
+    shipmode      VARCHAR(10) NOT NULL,
+    comment       VARCHAR(44) NOT NULL
+) WITH (format = 'PARQUET');
+
+-- Creating part table with PARQUET format
+CREATE TABLE part (
+    partkey     BIGINT NOT NULL,
+    name        VARCHAR(55) NOT NULL,
+    mfgr        VARCHAR(25) NOT NULL,
+    brand       VARCHAR(10) NOT NULL,
+    type        VARCHAR(25) NOT NULL,
+    size        BIGINT NOT NULL,
+    container   VARCHAR(10) NOT NULL,
+    retailprice DECIMAL(12,2) NOT NULL,
+    comment     VARCHAR(23) NOT NULL
+) WITH (format = 'PARQUET');
+
+-- Creating supplier table with PARQUET format
+CREATE TABLE supplier (
+    suppkey     BIGINT NOT NULL,
+    name        VARCHAR(25) NOT NULL,
+    address     VARCHAR(40) NOT NULL,
+    nationkey   BIGINT NOT NULL,
+    phone       VARCHAR(15) NOT NULL,
+    acctbal     DECIMAL(12,2) NOT NULL,
+    comment     VARCHAR(101) NOT NULL
+) WITH (format = 'PARQUET');
+
+-- Creating partsupp table with PARQUET format
+CREATE TABLE partsupp (
+    partkey    BIGINT NOT NULL,
+    suppkey    BIGINT NOT NULL,
+    availqty   BIGINT NOT NULL,
+    supplycost DECIMAL(12,2) NOT NULL,
+    comment    VARCHAR(199) NOT NULL
+) WITH (format = 'PARQUET');
+
+-- Creating nation table with PARQUET format
+CREATE TABLE nation (
+    nationkey   BIGINT NOT NULL,
+    name        VARCHAR(25) NOT NULL,
+    regionkey   BIGINT NOT NULL,
+    comment     VARCHAR(152) NOT NULL
+) WITH (format = 'PARQUET');
+
+-- Creating region table with PARQUET format
+CREATE TABLE region (
+    regionkey   BIGINT NOT NULL,
+    name        VARCHAR(25) NOT NULL,
+    comment     VARCHAR(152) NOT NULL
+) WITH (format = 'PARQUET');

benchmarks/insert/hive-tpch-parquet/lineitem_sf10.sql

@@ -0,0 +1,21 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into lineitem
+INSERT INTO lineitem
+SELECT
+    orderkey,
+    partkey,
+    suppkey,
+    linenumber,
+    CAST(quantity AS DECIMAL(12,2)),
+    CAST(extendedprice AS DECIMAL(12,2)),
+    CAST(discount AS DECIMAL(12,2)),
+    CAST(tax AS DECIMAL(12,2)),
+    CAST(returnflag AS VARCHAR(1)),
+    CAST(linestatus AS VARCHAR(1)),
+    shipdate,
+    commitdate,
+    receiptdate,
+    CAST(shipinstruct AS VARCHAR(25)),
+    CAST(shipmode AS VARCHAR(10)),
+    CAST(comment AS VARCHAR(44))
+FROM tpch.sf1.lineitem;

benchmarks/insert/hive-tpch-parquet/nation_sf10.sql

@@ -0,0 +1,9 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into nation
+INSERT INTO nation
+SELECT
+    nationkey,
+    CAST(name AS VARCHAR(25)),
+    regionkey,
+    CAST(comment AS VARCHAR(152))
+FROM tpch.sf1.nation;

benchmarks/insert/hive-tpch-parquet/orders_sf10.sql

@@ -0,0 +1,14 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into orders
+INSERT INTO orders
+SELECT
+    orderkey,
+    custkey,
+    CAST(orderstatus AS VARCHAR(1)),
+    CAST(totalprice AS DECIMAL(12,2)),
+    orderdate,
+    CAST(orderpriority AS VARCHAR(15)),
+    CAST(clerk AS VARCHAR(15)),
+    shippriority,
+    CAST(comment AS VARCHAR(79))
+FROM tpch.sf1.orders;

benchmarks/insert/hive-tpch-parquet/part_sf10.sql

@@ -0,0 +1,14 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into part
+INSERT INTO part
+SELECT
+    partkey,
+    CAST(name AS VARCHAR(55)),
+    CAST(mfgr AS VARCHAR(25)),
+    CAST(brand AS VARCHAR(10)),
+    CAST(type AS VARCHAR(25)),
+    size,
+    CAST(container AS VARCHAR(10)),
+    CAST(retailprice AS DECIMAL(12,2)),
+    CAST(comment AS VARCHAR(23))
+FROM tpch.sf1.part;

benchmarks/insert/hive-tpch-parquet/partsupp_sf10.sql

@@ -0,0 +1,10 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into partsupp
+INSERT INTO partsupp
+SELECT
+    partkey,
+    suppkey,
+    availqty,
+    CAST(supplycost AS DECIMAL(12,2)),
+    CAST(comment AS VARCHAR(199))
+FROM tpch.sf1.partsupp;

benchmarks/insert/hive-tpch-parquet/region_sf10.sql

@@ -0,0 +1,8 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into region
+INSERT INTO region
+SELECT
+    regionkey,
+    CAST(name AS VARCHAR(25)),
+    CAST(comment AS VARCHAR(152))
+FROM tpch.sf1.region;

benchmarks/insert/hive-tpch-parquet/supplier_sf10.sql

@@ -0,0 +1,12 @@
+USE tpch_sf10_load_iceberg;
+-- Inserting into supplier
+INSERT INTO supplier
+SELECT
+    suppkey,
+    CAST(name AS VARCHAR(25)),
+    CAST(address AS VARCHAR(40)),
+    nationkey,
+    CAST(phone AS VARCHAR(15)),
+    CAST(acctbal AS DECIMAL(12,2)),
+    CAST(comment AS VARCHAR(101))
+FROM tpch.sf1.supplier;