Test/hive dml (#155)

* feat: add showIndex parser rule

* test: uncomment show index test cases

* test: add unit tests about DML syntax to HiveSQL

* test: add unit tests about export and import syntax to HiveSQL

* refactor: recompile hive grammar

* test: correct description of HiveSQL unit tests

src/grammar/hive/HiveSqlParser.g4

@@ -346,6 +346,7 @@ showStatement
     | KW_SHOW KW_CONF StringLiteral
     | KW_SHOW KW_RESOURCE (KW_PLAN rp_name=id_ | KW_PLANS)
     | KW_SHOW KW_DATACONNECTORS
+    | KW_SHOW KW_FORMATTED? ( KW_INDEX | KW_INDEXES ) KW_ON tableName (from_in id_)?
     ;
 
 showTablesFilterExpr

src/lib/hive/HiveSqlLexer.ts

@@ -1,4 +1,4 @@
-// Generated from /Users/xuxiaoqi/Documents/dt-sql-parser-copy/src/grammar/hive/HiveSqlLexer.g4 by ANTLR 4.9.0-SNAPSHOT
+// Generated from /Users/hayden/Desktop/dt-works/dt-sql-parser/src/grammar/hive/HiveSqlLexer.g4 by ANTLR 4.9.0-SNAPSHOT
 
 
 import { ATN } from "antlr4ts/atn/ATN";

src/lib/hive/HiveSqlParserListener.ts

@@ -1,4 +1,4 @@
-// Generated from /Users/xuxiaoqi/Documents/dt-sql-parser-copy/src/grammar/hive/HiveSqlParser.g4 by ANTLR 4.9.0-SNAPSHOT
+// Generated from /Users/hayden/Desktop/dt-works/dt-sql-parser/src/grammar/hive/HiveSqlParser.g4 by ANTLR 4.9.0-SNAPSHOT
 
 
 import { ParseTreeListener } from "antlr4ts/tree/ParseTreeListener";

src/lib/hive/HiveSqlParserVisitor.ts

@@ -1,4 +1,4 @@
-// Generated from /Users/xuxiaoqi/Documents/dt-sql-parser-copy/src/grammar/hive/HiveSqlParser.g4 by ANTLR 4.9.0-SNAPSHOT
+// Generated from /Users/hayden/Desktop/dt-works/dt-sql-parser/src/grammar/hive/HiveSqlParser.g4 by ANTLR 4.9.0-SNAPSHOT
 
 
 import { ParseTreeVisitor } from "antlr4ts/tree/ParseTreeVisitor";

test/parser/hive/listener.test.ts

@@ -1,4 +1,5 @@
-import { HiveSqlListener } from '../../../src/lib/hive/HiveSqlListener';
+import { ProgramContext } from '../../../src/lib/hive/HiveSqlParser';
+import { HiveSqlParserListener } from '../../../src/lib/hive/HiveSqlParserListener';
 import HiveSQL from '../../../src/parser/hive';
 
 
@@ -10,28 +11,28 @@ describe('Hive SQL Listener Tests', () => {
         const parserTree = parser.parse(sql);
 
         let result = '';
-        class MyListener implements HiveSqlListener {
+        class MyListener implements HiveSqlParserListener {
             enterSelectItem(ctx) {
                 result = ctx.text;
             }
         }
         const listenTableName: any = new MyListener();
 
-        await parser.listen(listenTableName, parserTree);
+        await parser.listen(listenTableName, parserTree as ProgramContext);
         expect(result).toBe(expectTableName.toUpperCase());
     });
     test('Listener enterCreateTable', async () => {
         const sql = `drop table table_name;`;
         const parserTree = parser.parse(sql);
         let result = '';
-        class MyListener implements HiveSqlListener {
+        class MyListener implements HiveSqlParserListener {
             enterDropTableStatement(ctx) {
                 result = ctx.text;
             }
         }
         const listenTableName: any = new MyListener();
 
-        await parser.listen(listenTableName, parserTree);
+        await parser.listen(listenTableName, parserTree as ProgramContext);
         expect(result).toBe('DROPTABLETABLE_NAME');
     });
 });

test/parser/hive/syntax/abortStatement.test.ts

@@ -7,7 +7,7 @@ const features = {
     aborts: readSQL(__dirname, 'abort.sql'),
 };
 
-describe('Hive Abort Syntax Tests', () => {
+describe('HiveSQL Abort Syntax Tests', () => {
     features.aborts.forEach((ab) => {
         it(ab, () => {
             expect(parser.validate(ab).length).toBe(0);

test/parser/hive/syntax/alterStatement.test.ts

@@ -12,7 +12,7 @@ const features = {
     scheduleQueries: readSQL(__dirname, 'alterScheduleQuery.sql'),
 };
 
-describe('Hive Alter Syntax Tests', () => {
+describe('HiveSQL Alter Syntax Tests', () => {
     describe('ALTER DATABASE', () => {
         features.databases.forEach((db) => {
             it(db, () => {

test/parser/hive/syntax/createStatement.test.ts

@@ -15,7 +15,7 @@ const features = {
     scheduledQueries: readSQL(__dirname, 'createScheduledQuery.sql')
 };
 
-describe('Hive Create Syntax Tests', () => {
+describe('HiveSQL Create Syntax Tests', () => {
     describe('CREATE DATABASE', () => {
         features.databases.forEach((database) => {
             it(database, () => {

test/parser/hive/syntax/delete.test.ts

@@ -0,0 +1,16 @@
+import HiveSQL from '../../../../src/parser/hive';
+import { readSQL } from '../../../helper';
+
+const parser = new HiveSQL();
+
+const features = {
+    deletes: readSQL(__dirname, 'delete.sql'),
+};
+
+describe('HiveSQL Delete Syntax Tests', () => {
+    features.deletes.forEach((del) => {
+        it(del, () => {
+            expect(parser.validate(del).length).toBe(0);
+        });
+    });
+});

test/parser/hive/syntax/describeStatement.test.ts

@@ -7,7 +7,7 @@ const features = {
     desc: readSQL(__dirname, 'describe.sql'),
 };
 
-describe('Hive Describe Syntax Tests', () => {
+describe('HiveSQL Describe Syntax Tests', () => {
     features.desc.forEach((des) => {
         it(des, () => {
             expect(parser.validate(des).length).toBe(0);

test/parser/hive/syntax/dropStatement.test.ts

@@ -8,7 +8,7 @@ const features = {
     reloads: readSQL(__dirname, 'reload.sql')
 };
 
-describe('Hive Drop Syntax Tests', () => {
+describe('HiveSQL Drop Syntax Tests', () => {
     features.drops.forEach((drop) => {
         it(drop, () => {
             expect(parser.validate(drop).length).toBe(0);

test/parser/hive/syntax/exportStatement.test.ts

@@ -0,0 +1,17 @@
+import HiveSQL from '../../../../src/parser/hive';
+import { readSQL } from '../../../helper';
+
+const features = {
+    exports: readSQL(__dirname, 'export.sql')
+};
+
+describe('HiveSQL Export Syntax Tests', () => {
+    const parser = new HiveSQL();
+
+    features.exports.forEach((exp) => {
+        it(exp, () => {
+            expect(parser.validate(exp).length).toBe(0);
+        });
+    });
+});
+

test/parser/hive/syntax/fixtures/delete.sql

@@ -0,0 +1,3 @@
+DELETE FROM tbl;
+
+DELETE FROM tbl1 WHERE id = 1;

test/parser/hive/syntax/fixtures/export.sql

@@ -0,0 +1,6 @@
+EXPORT TABLE tbl TO 'hdfs_exports_location/department';
+
+EXPORT TABLE employee 
+PARTITION(emp_country="in", emp_state="ka") 
+TO 'hdfs_exports_location/employee'
+FOR REPLICATION('eventid');

test/parser/hive/syntax/fixtures/import.sql

@@ -0,0 +1,8 @@
+IMPORT FROM 'hdfs_exports_location/department';
+
+IMPORT TABLE employee PARTITION (emp_country="us", emp_state="tn") FROM 'hdfs_exports_location/employee';
+
+IMPORT EXTERNAL TABLE employee 
+PARTITION (emp_country="us", emp_state="tn") 
+FROM 'hdfs_exports_location/employee'
+LOCATION 'import_target_path';

test/parser/hive/syntax/fixtures/insertFormValues.sql

@@ -0,0 +1,11 @@
+INSERT INTO TABLE students
+VALUES ('fred flintstone', 35, 1.28), ('barney rubble', 32, 2.32);
+
+INSERT INTO TABLE pageviews PARTITION (datestamp = '2014-09-23')
+  VALUES ('jsmith', 'mail.com', 'sports.com'), ('jdoe', 'mail.com', null);
+ 
+INSERT INTO TABLE pageviews PARTITION (datestamp)
+  VALUES ('tjohnson', 'sports.com', 'finance.com', '2014-09-23'), ('tlee', 'finance.com', null, '2014-09-21');
+  
+INSERT INTO TABLE pageviews
+  VALUES ('tjohnson', 'sports.com', 'finance.com', '2014-09-23'), ('tlee', 'finance.com', null, '2014-09-21');

test/parser/hive/syntax/fixtures/insertFromQuery.sql

@@ -0,0 +1,79 @@
+-- Inserting data into Hive Tables from queries
+INSERT INTO table_name PARTITION (partition_col)
+SELECT col1, col2, partition_col
+FROM source_table;
+
+FROM source_table
+INSERT OVERWRITE TABLE table_name PARTITION (partition_col='value1')
+SELECT col1, col2
+WHERE partition_col = 'value1'
+INSERT INTO TABLE table_name PARTITION (partition_col='value2')
+SELECT col1, col2
+WHERE partition_col = 'value2';
+
+FROM page_view_stg pvs
+INSERT OVERWRITE TABLE page_view PARTITION(dt='2008-06-08', country)
+SELECT pvs.viewTime, pvs.userid, pvs.page_url, pvs.referrer_url, null, null, pvs.ip, pvs.cnt;
+
+
+-- Writing data into the filesystem from queries
+INSERT OVERWRITE LOCAL DIRECTORY '/path/to/output'
+SELECT col1, col2
+FROM table_name;
+
+INSERT OVERWRITE DIRECTORY '/path/to/output'
+STORED AS PARQUET
+SELECT col1, col2
+FROM table_name;
+
+INSERT INTO table_name PARTITION (year, month, day)
+SELECT col1, col2, 
+    CASE 
+        WHEN month = 'January' THEN 2023
+        WHEN month = 'February' THEN 2023
+        ELSE 2024
+    END AS year,
+    CASE 
+        WHEN month = 'January' THEN 1
+        WHEN month = 'February' THEN 2
+        ELSE 3
+    END AS month,
+    CAST(day AS int) AS day
+FROM source_table;
+
+INSERT INTO table_name PARTITION (country, state)
+SELECT col1, col2,
+    CONCAT(country, '_', state) AS country_state
+FROM source_table;
+
+INSERT INTO table_name PARTITION (country, state)
+SELECT col1, col2, country, state
+FROM (
+    SELECT col1, col2,
+        CASE
+            WHEN country = 'USA' THEN 'United States'
+            ELSE country
+        END AS country,
+        CASE
+            WHEN country = 'USA' THEN 'NA'
+            ELSE state
+        END AS state
+    FROM source_table
+) subquery;
+
+INSERT OVERWRITE LOCAL DIRECTORY '/path/to/output'
+ROW FORMAT DELIMITED
+FIELDS TERMINATED BY ',' 
+ESCAPED BY '^'
+COLLECTION ITEMS TERMINATED BY '.'
+MAP KEYS TERMINATED BY ':'
+LINES TERMINATED BY 'n'
+NULL DEFINED AS 'x'
+SELECT col1, col2
+FROM table_name;
+
+INSERT OVERWRITE LOCAL DIRECTORY '/path/to/output'
+ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.avro.AvroSerDe'
+STORED AS AVRO
+SELECT col1, col2
+FROM table_name;

test/parser/hive/syntax/fixtures/load.sql

@@ -0,0 +1,7 @@
+LOAD DATA INPATH 'hdfs://namenode:9000/user/hive/project/data1' INTO TABLE tbl1;
+
+LOAD DATA LOCAL INPATH '/user/hive/project/data1' 
+OVERWRITE INTO TABLE tablename 
+PARTITION (pt1=1, pt2=2)
+INPUTFORMAT 'com.apache.hadoop.textInputFormat'
+SERDE 'JsonSerDe';

test/parser/hive/syntax/fixtures/merge.sql

@@ -0,0 +1,10 @@
+MERGE INTO demo.a AS T1 USING demo.b AS T2
+ON T1.id = T2.id
+WHEN MATCHED THEN UPDATE SET name = 'wf1'
+WHEN NOT MATCHED THEN INSERT VALUES(T2.id,T2.name);
+
+MERGE INTO demo.a AS T1 USING demo.b AS T2
+ON T1.id = T2.id
+WHEN MATCHED AND sex='male' THEN UPDATE SET name = 'wf1'
+WHEN MATCHED AND sex='female' THEN UPDATE SET age = 10
+WHEN NOT MATCHED AND age>10 THEN INSERT VALUES(T2.id,T2.name);

test/parser/hive/syntax/fixtures/show.sql

@@ -39,9 +39,9 @@ SHOW TBLPROPERTIES tblname("foo");
 SHOW CREATE TABLE db.tbl1;
 
 -- Show Indexes
--- SHOW INDEX ON idx_tbl;
+SHOW INDEX ON idx_tbl;
 
--- SHOW FORMATTED INDEXES ON idx_tbl2 FROM db_1;
+SHOW FORMATTED INDEXES ON idx_tbl2 FROM db_1;
 
 -- Show Columns
 SHOW COLUMNS FROM tble;

test/parser/hive/syntax/fixtures/update.sql

@@ -0,0 +1,7 @@
+UPDATE table_name
+SET col1 = new_value;
+
+UPDATE table_name
+SET col1 = new_value,
+col2 = new_value2
+WHERE id=1;

test/parser/hive/syntax/importStatement.test.ts

@@ -0,0 +1,16 @@
+import HiveSQL from '../../../../src/parser/hive';
+import { readSQL } from "../../../helper";
+
+const features = {
+    imports: readSQL(__dirname, 'import.sql')
+};
+
+describe('HiveSQL Import Syntax Tests', () => {
+    const parser = new HiveSQL();
+
+    features.imports.forEach((imp) => {
+        it(imp, () => {
+            expect(parser.validate(imp).length).toBe(0);
+        });
+    });
+});

test/parser/hive/syntax/insertStatement.test.ts

@@ -0,0 +1,23 @@
+import HiveSQL from '../../../../src/parser/hive';
+import { readSQL } from '../../../helper';
+
+const parser = new HiveSQL();
+
+const features = {
+    insertFromQueries: readSQL(__dirname, 'insertFromQuery.sql'),
+    insertFromValues: readSQL(__dirname, 'insertFormValues.sql')
+};
+
+describe('HiveSQL Insert Syntax Tests', () => {
+    features.insertFromQueries.forEach((ifq) => {
+        it(ifq, () => {
+            expect(parser.validate(ifq).length).toBe(0);
+        });
+    });
+
+    features.insertFromValues.forEach((ifv) => {
+        it(ifv, () => {
+            expect(parser.validate(ifv).length).toBe(0);
+        });
+    });
+});

test/parser/hive/syntax/loadStatement.test.ts

@@ -0,0 +1,16 @@
+import HiveSQL from '../../../../src/parser/hive';
+import { readSQL } from '../../../helper';
+
+const parser = new HiveSQL();
+
+const features = {
+    loads: readSQL(__dirname, 'load.sql'),
+};
+
+describe('HiveSQL Load Syntax Tests', () => {
+    features.loads.forEach((load) => {
+        it(load, () => {
+            expect(parser.validate(load).length).toBe(0);
+        });
+    });
+});

test/parser/hive/syntax/merge.test.ts

@@ -0,0 +1,16 @@
+import HiveSQL from '../../../../src/parser/hive';
+import { readSQL } from '../../../helper';
+
+const parser = new HiveSQL();
+
+const features = {
+    merges: readSQL(__dirname, 'merge.sql'),
+};
+
+describe('HiveSQL Merge Syntax Tests', () => {
+    features.merges.forEach((merge) => {
+        it(merge, () => {
+            expect(parser.validate(merge).length).toBe(0);
+        });
+    });
+});

test/parser/hive/syntax/showStatement.test.ts

@@ -7,7 +7,7 @@ const features = {
     shows: readSQL(__dirname, 'show.sql'),
 };
 
-describe('Hive Show Syntax Tests', () => {
+describe('HiveSQL Show Syntax Tests', () => {
     features.shows.forEach((show) => {
         it(show, () => {
             expect(parser.validate(show).length).toBe(0);

test/parser/hive/syntax/update.test.ts

@@ -0,0 +1,16 @@
+import HiveSQL from '../../../../src/parser/hive';
+import { readSQL } from '../../../helper';
+
+const parser = new HiveSQL();
+
+const features = {
+    updates: readSQL(__dirname, 'update.sql'),
+};
+
+describe('HiveSQL Update Syntax Tests', () => {
+    features.updates.forEach((update) => {
+        it(update, () => {
+            expect(parser.validate(update).length).toBe(0);
+        });
+    });
+});

test/parser/hive/visitor.test.ts

@@ -1,6 +1,7 @@
 import { AbstractParseTreeVisitor } from 'antlr4ts/tree/AbstractParseTreeVisitor';
-import { HiveSqlVisitor } from '../../../src/lib/hive/HiveSqlVisitor';
+import { HiveSqlParserVisitor } from '../../../src/lib/hive/HiveSqlParserVisitor';
 import HiveSQL from '../../../src/parser/hive';
+import { ProgramContext } from '../../../src/lib/hive/HiveSqlParser';
 
 describe('HiveSQL Visitor Tests', () => {
     const expectTableName = 'dm_gis.dlv_addr_tc_count';
@@ -13,7 +14,7 @@ describe('Hive SQL Visitor Tests', () => {
 
     test('Visitor visitTableName', () => {
         let result = '';
-        class MyVisitor extends AbstractParseTreeVisitor<any> implements HiveSqlVisitor<any> {
+        class MyVisitor extends AbstractParseTreeVisitor<any> implements HiveSqlParserVisitor<any> {
 
             defaultResult() {
                 return result;
@@ -25,7 +26,7 @@ describe('Hive SQL Visitor Tests', () => {
         }
 
         const visitor = new MyVisitor();
-        visitor.visit(parserTree);
+        visitor.visit(parserTree as ProgramContext);
 
         expect(result).toBe(expectTableName);
     });