[#2569] improvement(trino): Optimize char/varchar mapping for Iceberg…

… catalog between Gravitino and Trino (#2579)

### What changes were proposed in this pull request?

Changing the char/varchar conversion for Iceberg catalog between
Gravitino and Trino.

### Why are the changes needed?

We need to make type conversion more exactly.

Fix: #2569 

### Does this PR introduce _any_ user-facing change?

N/A.

### How was this patch tested?

UT and ITs

integration-test/src/test/resources/trino-ci-testset/testsets/lakehouse-iceberg/00006_datatype.sql

@@ -4,8 +4,7 @@ USE "test.gt_iceberg".gt_db2;
 
 -- Unsupported Type: TINYINT, SMALLINT
 CREATE TABLE tb01 (
-    f1 VARCHAR(200),
-    f2 CHAR(20),
+    f1 VARCHAR,
     f3 VARBINARY,
     f4 DECIMAL(10, 3),
     f5 REAL,
@@ -21,17 +20,16 @@ CREATE TABLE tb01 (
 
 SHOW CREATE TABLE tb01;
 
-INSERT INTO tb01 (f1, f2, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
-VALUES ('Sample text 1', 'Text1', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, 100000, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
+INSERT INTO tb01 (f1, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
+VALUES ('Sample text 1', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, 100000, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
 
-INSERT INTO tb01 (f1, f2, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
-VALUES (NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL);
+INSERT INTO tb01 (f1, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
+VALUES (NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL);
 
 select * from tb01 order by f1;
 
 CREATE TABLE tb02 (
-    f1 VARCHAR(200) not null ,
-    f2 CHAR(20) not null ,
+    f1 VARCHAR not null ,
     f3 VARBINARY not null ,
     f4 DECIMAL(10, 3) not null ,
     f5 REAL not null ,
@@ -47,20 +45,20 @@ CREATE TABLE tb02 (
 
 show create table tb02;
 
-INSERT INTO tb02 (f1, f2, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
-VALUES ('Sample text 1', 'Text1', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, 100000, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
+INSERT INTO tb02 (f1, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
+VALUES ('Sample text 1', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, 100000, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
 
-INSERT INTO tb02 (f1, f2, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
-VALUES (NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL);
+INSERT INTO tb02 (f1, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
+VALUES (NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL);
 
-INSERT INTO tb02 (f1, f2, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
-VALUES ('Sample text 1', NULL, x'65', 123.456, 7.89, 12.34, true, 1000, 1000, 100000, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
+INSERT INTO tb02 (f1, f3, f4, f5, f6, f10, f11, f12, f13, f14, f15)
+VALUES ('Sample text 1', x'65', 123.456, 7.89, 12.34, 1000, 1000, 100000, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
 
-INSERT INTO tb02 (f1, f2, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
-VALUES ('Sample text 1', 'same3', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, NULl, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
+INSERT INTO tb02 (f1, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
+VALUES ('Sample text 1', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, NULL, DATE '2024-01-01', TIME '08:00:00', TIMESTAMP '2024-01-01 08:00:00');
 
-INSERT INTO tb02 (f1, f2, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
-VALUES ('Sample text 1', 'same9', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, 1992382342, DATE '2024-01-01', NULL, TIMESTAMP '2024-01-01 08:00:00');
+INSERT INTO tb02 (f1, f3, f4, f5, f6, f7, f10, f11, f12, f13, f14, f15)
+VALUES ('Sample text 1', x'65', 123.456, 7.89, 12.34, true, 1000, 1000, 1992382342, DATE '2024-01-01', NULL, TIMESTAMP '2024-01-01 08:00:00');
 
 drop table tb01;
 

integration-test/src/test/resources/trino-ci-testset/testsets/lakehouse-iceberg/00006_datatype.txt

@@ -6,7 +6,6 @@ CREATE TABLE
 
 "CREATE TABLE ""test.gt_iceberg"".gt_db2.tb01 (
    f1 varchar,
-   f2 varchar,
    f3 varbinary,
    f4 decimal(10, 3),
    f5 real,
@@ -26,14 +25,13 @@ INSERT: 1 row
 
 INSERT: 1 row
 
-"Sample text 1","Text1","65","123.456","7.89","12.34","true","1000","1000","100000","2024-01-01","08:00:00.000","2024-01-01 08:00:00.000"
-"","","","","","","","","","","","",""
+"Sample text 1","65","123.456","7.89","12.34","true","1000","1000","100000","2024-01-01","08:00:00.000","2024-01-01 08:00:00.000"
+"","","","","","","","","","","",""
 
 CREATE TABLE
 
 "CREATE TABLE ""test.gt_iceberg"".gt_db2.tb02 (
    f1 varchar NOT NULL,
-   f2 varchar NOT NULL,
    f3 varbinary NOT NULL,
    f4 decimal(10, 3) NOT NULL,
    f5 real NOT NULL,
@@ -52,7 +50,7 @@ INSERT: 1 row
 
 <QUERY_FAILED> NULL value not allowed for NOT NULL column: f1
 
-<QUERY_FAILED> NULL value not allowed for NOT NULL column: f2
+<QUERY_FAILED> NULL value not allowed for NOT NULL column: f7
 
 <QUERY_FAILED> NULL value not allowed for NOT NULL column: f12
 

integration-test/src/test/resources/trino-ci-testset/testsets/lakehouse-iceberg/00007_varchar.sql

@@ -0,0 +1,18 @@
+CREATE SCHEMA "test.gt_iceberg".varchar_db1;
+
+USE "test.gt_iceberg".varchar_db1;
+
+CREATE TABLE tb01 (id int, name char(20));
+
+CREATE TABLE tb02 (id int, name char);
+
+CREATE TABLE tb03 (id int, name varchar(233));
+
+CREATE TABLE tb04 (id int, name varchar);
+
+SHOW CREATE TABLE "test.gt_iceberg".varchar_db1.tb04;
+
+drop table "test.gt_iceberg".varchar_db1.tb04;
+
+drop schema "test.gt_iceberg".varchar_db1;
+

integration-test/src/test/resources/trino-ci-testset/testsets/lakehouse-iceberg/00007_varchar.txt

@@ -0,0 +1,21 @@
+CREATE SCHEMA
+
+USE
+
+<QUERY_FAILED> Iceberg does not support the datatype CHAR
+
+<QUERY_FAILED> Iceberg does not support the datatype CHAR
+
+<QUERY_FAILED> Iceberg does not support the datatype VARCHAR with length
+
+CREATE TABLE
+
+"CREATE TABLE ""test.gt_iceberg"".varchar_db1.tb04 (
+   id integer,
+   name varchar
+)
+COMMENT ''"
+
+DROP TABLE
+
+DROP SCHEMA

integration-test/src/test/resources/trino-ci-testset/testsets/tpch/catalog_iceberg_prepare.sql

@@ -12,13 +12,13 @@ use "test.gt_iceberg2".gt_tpch2;
 
 CREATE TABLE customer (
    custkey bigint,
-   name varchar(25),
-   address varchar(40),
+   name varchar,
+   address varchar,
    nationkey bigint,
-   phone varchar(15),
+   phone varchar,
    acctbal decimal(12, 2),
-   mktsegment varchar(10),
-   comment varchar(117)
+   mktsegment varchar,
+   comment varchar
 );
 
 CREATE TABLE lineitem (
@@ -30,69 +30,69 @@ CREATE TABLE lineitem (
    extendedprice decimal(12, 2),
    discount decimal(12, 2),
    tax decimal(12, 2),
-   returnflag varchar(1),
-   linestatus varchar(1),
+   returnflag varchar,
+   linestatus varchar,
    shipdate date,
    commitdate date,
    receiptdate date,
-   shipinstruct varchar(25),
-   shipmode varchar(10),
-   comment varchar(44)
+   shipinstruct varchar,
+   shipmode varchar,
+   comment varchar
 );
 
 CREATE TABLE nation (
    nationkey bigint,
-   name varchar(25),
+   name varchar,
    regionkey bigint,
-   comment varchar(152)
+   comment varchar
 );
 
 CREATE TABLE orders (
    orderkey bigint,
    custkey bigint,
-   orderstatus varchar(1),
+   orderstatus varchar,
    totalprice decimal(12, 2),
    orderdate date,
-   orderpriority varchar(15),
-   clerk varchar(15),
+   orderpriority varchar,
+   clerk varchar,
    shippriority integer,
-   comment varchar(79)
+   comment varchar
 );
 
 CREATE TABLE part (
    partkey bigint,
-   name varchar(55),
-   mfgr varchar(25),
-   brand varchar(10),
-   type varchar(25),
+   name varchar,
+   mfgr varchar,
+   brand varchar,
+   type varchar,
    size integer,
-   container varchar(10),
+   container varchar,
    retailprice decimal(12, 2),
-   comment varchar(23)
+   comment varchar
 );
 
 CREATE TABLE partsupp (
    partkey bigint,
    suppkey bigint,
    availqty integer,
    supplycost decimal(12, 2),
-   comment varchar(199)
+   comment varchar
 );
 
 CREATE TABLE region (
    regionkey bigint,
-   name varchar(25),
-   comment varchar(152)
+   name varchar,
+   comment varchar
 );
 
 CREATE TABLE supplier (
    suppkey bigint,
-   name varchar(25),
-   address varchar(40),
+   name varchar,
+   address varchar,
    nationkey bigint,
-   phone varchar(15),
+   phone varchar,
    acctbal decimal(12, 2),
-   comment varchar(101)
+   comment varchar
 );
 
 insert into customer select * from tpch.tiny.customer;

trino-connector/src/main/java/com/datastrato/gravitino/trino/connector/catalog/iceberg/IcebergDataTypeTransformer.java

@@ -8,19 +8,34 @@
 import com.datastrato.gravitino.rel.types.Type;
 import com.datastrato.gravitino.rel.types.Type.Name;
 import com.datastrato.gravitino.rel.types.Types;
+import com.datastrato.gravitino.trino.connector.GravitinoErrorCode;
 import com.datastrato.gravitino.trino.connector.util.GeneralDataTypeTransformer;
+import io.trino.spi.TrinoException;
 import io.trino.spi.type.VarbinaryType;
+import io.trino.spi.type.VarcharType;
 
 /** Type transformer between Iceberg and Trino */
 public class IcebergDataTypeTransformer extends GeneralDataTypeTransformer {
 
   @Override
   public Type getGravitinoType(io.trino.spi.type.Type type) {
-    Type gravitinoType = super.getGravitinoType(type);
-    if (gravitinoType.name() == Name.VARCHAR || gravitinoType.name() == Name.FIXEDCHAR) {
+    Class<? extends io.trino.spi.type.Type> typeClass = type.getClass();
+    if (typeClass == io.trino.spi.type.CharType.class) {
+      throw new TrinoException(
+          GravitinoErrorCode.GRAVITINO_ILLEGAL_ARGUMENT,
+          "Iceberg does not support the datatype CHAR");
+    } else if (typeClass == io.trino.spi.type.VarcharType.class) {
+      VarcharType varCharType = (VarcharType) type;
+      if (varCharType.getLength().isPresent()) {
+        throw new TrinoException(
+            GravitinoErrorCode.GRAVITINO_ILLEGAL_ARGUMENT,
+            "Iceberg does not support the datatype VARCHAR with length");
+      }
+
       return Types.StringType.get();
     }
-    return gravitinoType;
+
+    return super.getGravitinoType(type);
   }
 
   @Override

trino-connector/src/test/java/com/datastrato/gravitino/trino/connector/catalog/iceberg/TestIcebergDataTypeTransformer.java

@@ -0,0 +1,41 @@
+/*
+ * Copyright 2024 Datastrato Pvt Ltd.
+ * This software is licensed under the Apache License version 2.
+ */
+
+package com.datastrato.gravitino.trino.connector.catalog.iceberg;
+
+import com.datastrato.gravitino.rel.types.Types;
+import com.datastrato.gravitino.trino.connector.util.GeneralDataTypeTransformer;
+import io.trino.spi.TrinoException;
+import io.trino.spi.type.VarcharType;
+import org.testng.Assert;
+import org.testng.annotations.Test;
+
+public class TestIcebergDataTypeTransformer {
+
+  @Test
+  public void testTrinoTypeToGravitinoType() {
+    GeneralDataTypeTransformer generalDataTypeTransformer = new IcebergDataTypeTransformer();
+    io.trino.spi.type.Type charTypeWithLengthOne = io.trino.spi.type.CharType.createCharType(1);
+
+    Exception e =
+        Assert.expectThrows(
+            TrinoException.class,
+            () -> generalDataTypeTransformer.getGravitinoType(charTypeWithLengthOne));
+    Assert.assertTrue(e.getMessage().contains("Iceberg does not support the datatype CHAR"));
+
+    io.trino.spi.type.Type varcharType = io.trino.spi.type.VarcharType.createVarcharType(1);
+    e =
+        Assert.expectThrows(
+            TrinoException.class, () -> generalDataTypeTransformer.getGravitinoType(varcharType));
+    Assert.assertTrue(
+        e.getMessage().contains("Iceberg does not support the datatype VARCHAR with length"));
+
+    io.trino.spi.type.Type varcharTypeWithoutLength = VarcharType.VARCHAR;
+
+    Assert.assertEquals(
+        generalDataTypeTransformer.getGravitinoType(varcharTypeWithoutLength),
+        Types.StringType.get());
+  }
+}