huaxingao commented on a change in pull request #29972: URL: https://github.com/apache/spark/pull/29972#discussion_r504154239
########## File path: external/docker-integration-tests/src/test/scala/org/apache/spark/sql/jdbc/v2/V2JDBCTest.scala ########## @@ -0,0 +1,98 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.spark.sql.jdbc.v2 + +import org.apache.spark.sql.AnalysisException +import org.apache.spark.sql.test.SharedSparkSession +import org.apache.spark.sql.types._ +import org.apache.spark.tags.DockerTest + +@DockerTest +trait V2JDBCTest extends SharedSparkSession { + val catalogName: String + // dialect specific update column type test + def testUpdateColumnType(tbl: String): Unit + + test("SPARK-33034: ALTER TABLE ... add new columns") { + withTable(s"$catalogName.alt_table") { + sql(s"CREATE TABLE $catalogName.alt_table (ID STRING) USING _") + var t = spark.table(s"$catalogName.alt_table") + var expectedSchema = new StructType().add("ID", StringType) + assert(t.schema === expectedSchema) + sql(s"ALTER TABLE $catalogName.alt_table ADD COLUMNS (C1 STRING, C2 STRING)") + t = spark.table(s"$catalogName.alt_table") + expectedSchema = expectedSchema.add("C1", StringType).add("C2", StringType) + assert(t.schema === expectedSchema) + sql(s"ALTER TABLE $catalogName.alt_table ADD COLUMNS (C3 STRING)") + t = spark.table(s"$catalogName.alt_table") + expectedSchema = expectedSchema.add("C3", StringType) + assert(t.schema === expectedSchema) + // Add already existing column + val msg = intercept[AnalysisException] { + sql(s"ALTER TABLE $catalogName.alt_table ADD COLUMNS (C3 DOUBLE)") + }.getMessage + assert(msg.contains("Cannot add column, because C3 already exists")) + } + // Add a column to not existing table + val msg = intercept[AnalysisException] { + sql(s"ALTER TABLE $catalogName.not_existing_table ADD COLUMNS (C4 STRING)") + }.getMessage + assert(msg.contains("Table not found")) + } + + test("SPARK-33034: ALTER TABLE ... update column type") { + withTable(s"$catalogName.alt_table") { + testUpdateColumnType(s"$catalogName.alt_table") + // Update not existing column + val msg2 = intercept[AnalysisException] { + sql(s"ALTER TABLE $catalogName.alt_table ALTER COLUMN bad_column TYPE DOUBLE") + }.getMessage + assert(msg2.contains("Cannot update missing field bad_column")) + } + // Update column type in not existing table + val msg = intercept[AnalysisException] { + sql(s"ALTER TABLE $catalogName.not_existing_table ALTER COLUMN id TYPE DOUBLE") + }.getMessage + assert(msg.contains("Table not found")) + } + + test("SPARK-33034: ALTER TABLE ... update column nullability") { + withTable(s"$catalogName.alt_table") { + sql(s"CREATE TABLE $catalogName.alt_table (ID STRING NOT NULL) USING _") + var t = spark.table(s"$catalogName.alt_table") + // nullable is true in the expecteSchema because Spark always sets nullable to true + // regardless of the JDBC metadata https://github.com/apache/spark/pull/18445 Review comment: I did a couple of quick tests using V2 write API: ``` sql("INSERT INTO h2.test.people SELECT 'bob', null") ``` and ``` sql("SELECT null AS ID, 'bob' AS NAME").writeTo("h2.test.people") ``` I got Exception from h2 jdbc driver: ``` Caused by: org.h2.jdbc.JdbcSQLException: NULL not allowed for column "ID"; SQL statement: INSERT INTO "test"."people" ("NAME","ID") VALUES (?,?) [23502-195] at org.h2.message.DbException.getJdbcSQLException(DbException.java:345) ``` So we are able to pass the null value for not null column `ID` to h2 and h2 blocks the insert. However, if I change the current code in `JDBCRDD.resolveTable` to make `alwaysNullable = false` to get the real nullable value, ``` def resolveTable(options: JDBCOptions): StructType = { ...... JdbcUtils.getSchema(rs, dialect, alwaysNullable = false) ``` For insert, I got Exception from Spark ``` Cannot write incompatible data to table 'test.people': - Cannot write nullable values to non-null column 'ID'; org.apache.spark.sql.AnalysisException: Cannot write incompatible data to table 'test.people': - Cannot write nullable values to non-null column 'ID'; at org.apache.spark.sql.catalyst.analysis.TableOutputResolver$.resolveOutputColumns(TableOutputResolver.scala:72) at org.apache.spark.sql.catalyst.analysis.Analyzer$ResolveOutputRelation$$anonfun$apply$31.applyOrElse(Analyzer.scala:3040) at org.apache.spark.sql.catalyst.analysis.Analyzer$ResolveOutputRelation$$anonfun$apply$31.applyOrElse(Analyzer.scala:3035) ``` Spark blocks the insert and we are not able to pass the null value for not null column ID to h2. Since the whole point of https://github.com/apache/spark/pull/18445 is to let the underlying database to decide how to process null for a not null column, I guess we will not change this `alwaysNullable` for JDBCV2? ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
