[
https://issues.apache.org/jira/browse/FLINK-24537?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Caizhi Weng updated FLINK-24537:
--------------------------------
Description:
This issue is from [the mailing
list|https://lists.apache.org/thread.html/r112786ef139a1029f13565852f66bf542ff9bd62394cedf5b471cac2%40%3Cuser-zh.flink.apache.org%3E].
Add the following test case to {{TableEnvironmentITCase}} to reproduce this
issue.
{code:scala}
@Test
def myTest(): Unit = {
tEnv.executeSql(
"""
|CREATE TABLE T (
| aa INT,
| b INT
|) WITH (
| 'connector' = 'values'
|)
|""".stripMargin)
tEnv.executeSql(
"""
|SELECT array[row('aa', aa), row('b', b)] FROM T
|""".stripMargin).print()
}
{code}
The exception stack is
{code:java}
java.lang.AssertionError: Conversion to relational algebra failed to preserve
datatypes:
validated type:
RecordType(RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT NULL EXPR$0,
INTEGER EXPR$1) NOT NULL ARRAY NOT NULL EXPR$0) NOT NULL
converted type:
RecordType(RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT NULL EXPR$0,
INTEGER NOT NULL EXPR$1) NOT NULL ARRAY NOT NULL EXPR$0) NOT NULL
rel:
LogicalProject(EXPR$0=[ARRAY(CAST(ROW(_UTF-16LE'aa', $0)):RecordType(VARCHAR(2)
CHARACTER SET "UTF-16LE" NOT NULL EXPR$0, INTEGER NOT NULL EXPR$1) NOT NULL,
CAST(ROW(_UTF-16LE'b', $1)):RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT
NULL EXPR$0, INTEGER NOT NULL EXPR$1) NOT NULL)])
LogicalTableScan(table=[[default_catalog, default_database, T]])
at
org.apache.calcite.sql2rel.SqlToRelConverter.checkConvertedType(SqlToRelConverter.java:467)
at
org.apache.calcite.sql2rel.SqlToRelConverter.convertQuery(SqlToRelConverter.java:582)
at
org.apache.flink.table.planner.calcite.FlinkPlannerImpl.org$apache$flink$table$planner$calcite$FlinkPlannerImpl$$rel(FlinkPlannerImpl.scala:181)
at
org.apache.flink.table.planner.calcite.FlinkPlannerImpl.rel(FlinkPlannerImpl.scala:173)
at
org.apache.flink.table.planner.operations.SqlToOperationConverter.toQueryOperation(SqlToOperationConverter.java:1074)
at
org.apache.flink.table.planner.operations.SqlToOperationConverter.convertSqlQuery(SqlToOperationConverter.java:1043)
at
org.apache.flink.table.planner.operations.SqlToOperationConverter.convert(SqlToOperationConverter.java:305)
at
org.apache.flink.table.planner.delegation.ParserImpl.parse(ParserImpl.java:101)
at
org.apache.flink.table.api.internal.TableEnvironmentImpl.executeSql(TableEnvironmentImpl.java:739)
{code}
One interesting aspect is that the following two SQLs are OK.
{code:sql}
-- constant strings are of same length
SELECT array[row('a', aa), row('b', b)] FROM T
-- constant strings are casted to varchar
SELECT array[row(CAST('aa' AS VARCHAR), aa), row(CAST('b' AS VARCHAR), b)] FROM
T
{code}
was:
Add the following test case to {{TableEnvironmentITCase}} to reproduce this
issue.
{code:scala}
@Test
def myTest(): Unit = {
tEnv.executeSql(
"""
|CREATE TABLE T (
| aa INT,
| b INT
|) WITH (
| 'connector' = 'values'
|)
|""".stripMargin)
tEnv.executeSql(
"""
|SELECT array[row('aa', aa), row('b', b)] FROM T
|""".stripMargin).print()
}
{code}
The exception stack is
{code}
java.lang.AssertionError: Conversion to relational algebra failed to preserve
datatypes:
validated type:
RecordType(RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT NULL EXPR$0,
INTEGER EXPR$1) NOT NULL ARRAY NOT NULL EXPR$0) NOT NULL
converted type:
RecordType(RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT NULL EXPR$0,
INTEGER NOT NULL EXPR$1) NOT NULL ARRAY NOT NULL EXPR$0) NOT NULL
rel:
LogicalProject(EXPR$0=[ARRAY(CAST(ROW(_UTF-16LE'aa', $0)):RecordType(VARCHAR(2)
CHARACTER SET "UTF-16LE" NOT NULL EXPR$0, INTEGER NOT NULL EXPR$1) NOT NULL,
CAST(ROW(_UTF-16LE'b', $1)):RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT
NULL EXPR$0, INTEGER NOT NULL EXPR$1) NOT NULL)])
LogicalTableScan(table=[[default_catalog, default_database, T]])
at
org.apache.calcite.sql2rel.SqlToRelConverter.checkConvertedType(SqlToRelConverter.java:467)
at
org.apache.calcite.sql2rel.SqlToRelConverter.convertQuery(SqlToRelConverter.java:582)
at
org.apache.flink.table.planner.calcite.FlinkPlannerImpl.org$apache$flink$table$planner$calcite$FlinkPlannerImpl$$rel(FlinkPlannerImpl.scala:181)
at
org.apache.flink.table.planner.calcite.FlinkPlannerImpl.rel(FlinkPlannerImpl.scala:173)
at
org.apache.flink.table.planner.operations.SqlToOperationConverter.toQueryOperation(SqlToOperationConverter.java:1074)
at
org.apache.flink.table.planner.operations.SqlToOperationConverter.convertSqlQuery(SqlToOperationConverter.java:1043)
at
org.apache.flink.table.planner.operations.SqlToOperationConverter.convert(SqlToOperationConverter.java:305)
at
org.apache.flink.table.planner.delegation.ParserImpl.parse(ParserImpl.java:101)
at
org.apache.flink.table.api.internal.TableEnvironmentImpl.executeSql(TableEnvironmentImpl.java:739)
{code}
One interesting aspect is that the following SQL is OK.
{code:sql}
-- constant strings are of same length
SELECT array[row('a', aa), row('b', b)] FROM T
{code}
> Array of rows containing constant strings of different length and an integer
> throws AssertionError
> --------------------------------------------------------------------------------------------------
>
> Key: FLINK-24537
> URL: https://issues.apache.org/jira/browse/FLINK-24537
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.15.0
> Reporter: Caizhi Weng
> Priority: Major
>
> This issue is from [the mailing
> list|https://lists.apache.org/thread.html/r112786ef139a1029f13565852f66bf542ff9bd62394cedf5b471cac2%40%3Cuser-zh.flink.apache.org%3E].
> Add the following test case to {{TableEnvironmentITCase}} to reproduce this
> issue.
> {code:scala}
> @Test
> def myTest(): Unit = {
> tEnv.executeSql(
> """
> |CREATE TABLE T (
> | aa INT,
> | b INT
> |) WITH (
> | 'connector' = 'values'
> |)
> |""".stripMargin)
> tEnv.executeSql(
> """
> |SELECT array[row('aa', aa), row('b', b)] FROM T
> |""".stripMargin).print()
> }
> {code}
> The exception stack is
> {code:java}
> java.lang.AssertionError: Conversion to relational algebra failed to preserve
> datatypes:
> validated type:
> RecordType(RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT NULL EXPR$0,
> INTEGER EXPR$1) NOT NULL ARRAY NOT NULL EXPR$0) NOT NULL
> converted type:
> RecordType(RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT NULL EXPR$0,
> INTEGER NOT NULL EXPR$1) NOT NULL ARRAY NOT NULL EXPR$0) NOT NULL
> rel:
> LogicalProject(EXPR$0=[ARRAY(CAST(ROW(_UTF-16LE'aa',
> $0)):RecordType(VARCHAR(2) CHARACTER SET "UTF-16LE" NOT NULL EXPR$0, INTEGER
> NOT NULL EXPR$1) NOT NULL, CAST(ROW(_UTF-16LE'b', $1)):RecordType(VARCHAR(2)
> CHARACTER SET "UTF-16LE" NOT NULL EXPR$0, INTEGER NOT NULL EXPR$1) NOT NULL)])
> LogicalTableScan(table=[[default_catalog, default_database, T]])
> at
> org.apache.calcite.sql2rel.SqlToRelConverter.checkConvertedType(SqlToRelConverter.java:467)
> at
> org.apache.calcite.sql2rel.SqlToRelConverter.convertQuery(SqlToRelConverter.java:582)
> at
> org.apache.flink.table.planner.calcite.FlinkPlannerImpl.org$apache$flink$table$planner$calcite$FlinkPlannerImpl$$rel(FlinkPlannerImpl.scala:181)
> at
> org.apache.flink.table.planner.calcite.FlinkPlannerImpl.rel(FlinkPlannerImpl.scala:173)
> at
> org.apache.flink.table.planner.operations.SqlToOperationConverter.toQueryOperation(SqlToOperationConverter.java:1074)
> at
> org.apache.flink.table.planner.operations.SqlToOperationConverter.convertSqlQuery(SqlToOperationConverter.java:1043)
> at
> org.apache.flink.table.planner.operations.SqlToOperationConverter.convert(SqlToOperationConverter.java:305)
> at
> org.apache.flink.table.planner.delegation.ParserImpl.parse(ParserImpl.java:101)
> at
> org.apache.flink.table.api.internal.TableEnvironmentImpl.executeSql(TableEnvironmentImpl.java:739)
> {code}
> One interesting aspect is that the following two SQLs are OK.
> {code:sql}
> -- constant strings are of same length
> SELECT array[row('a', aa), row('b', b)] FROM T
> -- constant strings are casted to varchar
> SELECT array[row(CAST('aa' AS VARCHAR), aa), row(CAST('b' AS VARCHAR), b)]
> FROM T
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
