[#2568] improvement(trino): Optimize char/varchar mapping for Hive ca…

…talog between Gravitino and Trino (#2576)

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

Changing the char/varchar conversion between Gravitino and Trino.

### Why are the changes needed?

We need to make type conversion more exactly. 

Fix: #2568 

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

N/A.

### How was this patch tested?

UT and IT

integration-test/src/test/resources/trino-ci-testset/testsets/hive/00002_alter_table.txt

@@ -13,7 +13,7 @@ RENAME TABLE
 DROP COLUMN
 
 "CREATE TABLE ""test.gt_hive"".gt_db1.tb01 (
-   name varchar(65535),
+   name varchar,
    salary integer
 )
 COMMENT ''
@@ -31,7 +31,7 @@ WITH (
 RENAME COLUMN
 
 "CREATE TABLE ""test.gt_hive"".gt_db1.tb01 (
-   s varchar(65535),
+   s varchar,
    salary integer
 )
 COMMENT ''
@@ -105,7 +105,7 @@ ADD COLUMN
 "CREATE TABLE ""test.gt_hive"".gt_db1.tb01 (
    s varchar(256) COMMENT 'test column comments',
    salary integer,
-   city varchar(65535) COMMENT 'aaa'
+   city varchar COMMENT 'aaa'
 )
 COMMENT 'test table comments'
 WITH (

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

@@ -0,0 +1,47 @@
+CREATE SCHEMA "test.gt_hive".varchar_db1;
+
+USE "test.gt_hive".varchar_db1;
+
+CREATE TABLE tb01 (id int, name char(20));
+
+SHOW CREATE TABLE "test.gt_hive".varchar_db1.tb01;
+
+CREATE TABLE tb02 (id int, name char(255));
+
+SHOW CREATE TABLE "test.gt_hive".varchar_db1.tb02;
+
+CREATE TABLE tb03 (id int, name char(256));
+
+CREATE TABLE tb04 (id int, name varchar(250));
+
+SHOW CREATE TABLE "test.gt_hive".varchar_db1.tb04;
+
+CREATE TABLE tb05 (id int, name varchar(65535));
+
+SHOW CREATE TABLE "test.gt_hive".varchar_db1.tb05;
+
+CREATE TABLE tb06 (id int, name char);
+
+SHOW CREATE TABLE "test.gt_hive".varchar_db1.tb06;
+
+CREATE TABLE tb07 (id int, name varchar);
+
+SHOW CREATE TABLE "test.gt_hive".varchar_db1.tb07;
+
+CREATE TABLE tb08 (id int, name varchar(65536));
+
+
+drop table "test.gt_hive".varchar_db1.tb01;
+
+drop table "test.gt_hive".varchar_db1.tb02;
+
+drop table "test.gt_hive".varchar_db1.tb04;
+
+drop table "test.gt_hive".varchar_db1.tb05;
+
+drop table "test.gt_hive".varchar_db1.tb06;
+
+drop table "test.gt_hive".varchar_db1.tb07;
+
+drop schema "test.gt_hive".varchar_db1;
+

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

@@ -0,0 +1,117 @@
+CREATE SCHEMA
+
+USE
+
+CREATE TABLE
+
+"CREATE TABLE ""test.gt_hive"".varchar_db1.tb01 (
+   id integer,
+   name char(20)
+)
+COMMENT ''
+WITH (
+   input_format = 'org.apache.hadoop.mapred.TextInputFormat',
+   location = 'hdfs://%:9000/user/hive/warehouse/varchar_db1.db/tb01',
+   output_format = 'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat',
+   serde_lib = 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe',
+   serde_name = 'tb01',
+   table_type = 'MANAGED_TABLE'
+)"
+
+CREATE TABLE
+
+"CREATE TABLE ""test.gt_hive"".varchar_db1.tb02 (
+   id integer,
+   name char(255)
+)
+COMMENT ''
+WITH (
+   input_format = 'org.apache.hadoop.mapred.TextInputFormat',
+   location = 'hdfs://%:9000/user/hive/warehouse/varchar_db1.db/tb02',
+   output_format = 'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat',
+   serde_lib = 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe',
+   serde_name = 'tb02',
+   table_type = 'MANAGED_TABLE'
+)"
+
+<QUERY_FAILED> Hive does not support the datatype CHAR with the length greater than 255
+
+CREATE TABLE
+
+"CREATE TABLE ""test.gt_hive"".varchar_db1.tb04 (
+   id integer,
+   name varchar(250)
+)
+COMMENT ''
+WITH (
+   input_format = 'org.apache.hadoop.mapred.TextInputFormat',
+   location = 'hdfs://%:9000/user/hive/warehouse/varchar_db1.db/tb04',
+   output_format = 'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat',
+   serde_lib = 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe',
+   serde_name = 'tb04',
+   table_type = 'MANAGED_TABLE'
+)"
+
+CREATE TABLE
+
+"CREATE TABLE ""test.gt_hive"".varchar_db1.tb05 (
+   id integer,
+   name varchar(65535)
+)
+COMMENT ''
+WITH (
+   input_format = 'org.apache.hadoop.mapred.TextInputFormat',
+   location = 'hdfs://%:9000/user/hive/warehouse/varchar_db1.db/tb05',
+   output_format = 'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat',
+   serde_lib = 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe',
+   serde_name = 'tb05',
+   table_type = 'MANAGED_TABLE'
+)"
+
+CREATE TABLE
+
+"CREATE TABLE ""test.gt_hive"".varchar_db1.tb06 (
+   id integer,
+   name char(1)
+)
+COMMENT ''
+WITH (
+   input_format = 'org.apache.hadoop.mapred.TextInputFormat',
+   location = 'hdfs://%:9000/user/hive/warehouse/varchar_db1.db/tb06',
+   output_format = 'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat',
+   serde_lib = 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe',
+   serde_name = 'tb06',
+   table_type = 'MANAGED_TABLE'
+)"
+
+CREATE TABLE
+
+"CREATE TABLE ""test.gt_hive"".varchar_db1.tb07 (
+   id integer,
+   name varchar
+)
+COMMENT ''
+WITH (
+   input_format = 'org.apache.hadoop.mapred.TextInputFormat',
+   location = 'hdfs://%:9000/user/hive/warehouse/varchar_db1.db/tb07',
+   output_format = 'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat',
+   serde_lib = 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe',
+   serde_name = 'tb07',
+   table_type = 'MANAGED_TABLE'
+)"
+
+<QUERY_FAILED> Hive does not support the datatype VARCHAR with the length greater than 65535, you can use varchar without length instead
+
+DROP TABLE
+
+DROP TABLE
+
+DROP TABLE
+
+DROP TABLE
+
+DROP TABLE
+
+DROP TABLE
+
+DROP SCHEMA

trino-connector/src/main/java/com/datastrato/gravitino/trino/connector/catalog/hive/HiveDataTypeTransformer.java

@@ -6,28 +6,49 @@
 package com.datastrato.gravitino.trino.connector.catalog.hive;
 
 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.VarcharType;
 
 /** Type transformer between Hive and Trino */
 public class HiveDataTypeTransformer extends GeneralDataTypeTransformer {
   // Max length of Hive varchar is 65535
   private static final int HIVE_VARCHAR_MAX_LENGTH = 65535;
+  private static final int HIVE_CHAR_MAX_LENGTH = 255;
 
   @Override
   public Type getGravitinoType(io.trino.spi.type.Type type) {
-    Type gravitinoType = super.getGravitinoType(type);
-    if (gravitinoType.name() == Name.VARCHAR
-        && ((Types.VarCharType) gravitinoType).length() > HIVE_VARCHAR_MAX_LENGTH) {
-      return Types.VarCharType.of(HIVE_VARCHAR_MAX_LENGTH);
-    }
+    Class<? extends io.trino.spi.type.Type> typeClass = type.getClass();
+    if (typeClass == VarcharType.class) {
+      VarcharType varcharType = (VarcharType) type;
+      if (varcharType.getLength().isEmpty()) {
+        return Types.StringType.get();
+      }
+
+      int length = varcharType.getLength().get();
+      if (length > HIVE_VARCHAR_MAX_LENGTH) {
+        throw new TrinoException(
+            GravitinoErrorCode.GRAVITINO_ILLEGAL_ARGUMENT,
+            "Hive does not support the datatype VARCHAR with the length greater than "
+                + HIVE_VARCHAR_MAX_LENGTH
+                + ", you can use varchar without length instead");
+      }
+
+      return Types.VarCharType.of(length);
+    } else if (typeClass == io.trino.spi.type.CharType.class) {
+      io.trino.spi.type.CharType charType = (io.trino.spi.type.CharType) type;
+      if (charType.getLength() > HIVE_CHAR_MAX_LENGTH) {
+        throw new TrinoException(
+            GravitinoErrorCode.GRAVITINO_ILLEGAL_ARGUMENT,
+            "Hive does not support the datatype CHAR with the length greater than "
+                + HIVE_CHAR_MAX_LENGTH);
+      }
 
-    if (gravitinoType.name() == Name.FIXEDCHAR
-        && ((Types.FixedCharType) gravitinoType).length() > HIVE_VARCHAR_MAX_LENGTH) {
-      return Types.FixedCharType.of(HIVE_VARCHAR_MAX_LENGTH);
+      return Types.FixedCharType.of(charType.getLength());
     }
 
-    return gravitinoType;
+    return super.getGravitinoType(type);
   }
 }

trino-connector/src/test/java/com/datastrato/gravitino/trino/connector/catalog/hive/TestHiveDataTypeConverter.java

@@ -0,0 +1,66 @@
+/*
+ * Copyright 2024 Datastrato Pvt Ltd.
+ * This software is licensed under the Apache License version 2.
+ */
+
+package com.datastrato.gravitino.trino.connector.catalog.hive;
+
+import com.datastrato.gravitino.rel.types.Types;
+import com.datastrato.gravitino.trino.connector.util.GeneralDataTypeTransformer;
+import io.trino.spi.TrinoException;
+import org.testng.Assert;
+import org.testng.annotations.Test;
+
+public class TestHiveDataTypeConverter {
+
+  @Test
+  public void testTrinoTypeToGravitinoType() {
+    GeneralDataTypeTransformer generalDataTypeTransformer = new HiveDataTypeTransformer();
+    io.trino.spi.type.Type charTypeWithLengthOne = io.trino.spi.type.CharType.createCharType(1);
+    Assert.assertEquals(
+        generalDataTypeTransformer.getGravitinoType(charTypeWithLengthOne),
+        Types.FixedCharType.of(1));
+
+    io.trino.spi.type.Type charTypeWithLength = io.trino.spi.type.CharType.createCharType(255);
+    Assert.assertEquals(
+        generalDataTypeTransformer.getGravitinoType(charTypeWithLength),
+        Types.FixedCharType.of(255));
+
+    io.trino.spi.type.Type charLengthIsOverflow = io.trino.spi.type.CharType.createCharType(256);
+    Exception e =
+        Assert.expectThrows(
+            TrinoException.class,
+            () -> generalDataTypeTransformer.getGravitinoType(charLengthIsOverflow));
+    Assert.assertTrue(
+        e.getMessage()
+            .contains("Hive does not support the datatype CHAR with the length greater than 255"));
+
+    io.trino.spi.type.Type varcharType = io.trino.spi.type.VarcharType.createVarcharType(1);
+    Assert.assertEquals(
+        generalDataTypeTransformer.getGravitinoType(varcharType), Types.VarCharType.of(1));
+
+    io.trino.spi.type.Type varcharTypeWithLength =
+        io.trino.spi.type.VarcharType.createVarcharType(65535);
+    Assert.assertEquals(
+        generalDataTypeTransformer.getGravitinoType(varcharTypeWithLength),
+        Types.VarCharType.of(65535));
+
+    io.trino.spi.type.Type varcharLengthIsOverflow =
+        io.trino.spi.type.VarcharType.createVarcharType(65536);
+    e =
+        Assert.expectThrows(
+            TrinoException.class,
+            () -> generalDataTypeTransformer.getGravitinoType(varcharLengthIsOverflow));
+    Assert.assertTrue(
+        e.getMessage()
+            .contains(
+                "Hive does not support the datatype VARCHAR with the length greater than 65535"));
+
+    io.trino.spi.type.Type varcharTypeWithoutLength =
+        io.trino.spi.type.VarcharType.createUnboundedVarcharType();
+
+    Assert.assertEquals(
+        generalDataTypeTransformer.getGravitinoType(varcharTypeWithoutLength),
+        Types.StringType.get());
+  }
+}