[
https://issues.apache.org/jira/browse/FLINK-38913?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18052013#comment-18052013
]
david radley commented on FLINK-38913:
--------------------------------------
The
[Grammar](https://github.com/apache/flink/blob/4def6efb47ebc2dff46f444a3319d3c2ae02c3a6/flink-table/flink-sql-parser/src/main/codegen/includes/parserImpls.ftl#L2588)shows
that the comments should have a null or text for every field. So I am not sure
how we can get into this out of bounds situation.
> ArrayIndexOutOfBoundsException when creating a table with computed rows
> including casts to null
> -----------------------------------------------------------------------------------------------
>
> Key: FLINK-38913
> URL: https://issues.apache.org/jira/browse/FLINK-38913
> Project: Flink
> Issue Type: Technical Debt
> Components: Table SQL / Planner
> Affects Versions: 2.2.0
> Reporter: david radley
> Assignee: david radley
> Priority: Major
>
> If you try to create a table in the SQL client with SQL like this :
> {code:sql}
> CREATE TABLE orders (
> `const_requestBody_arrayOfObjectsOfObjects` AS ARRAY[ROW(ROW('I''m
> happy', CAST(NULL AS BOOLEAN))), ROW(CAST(NULL AS ROW<`string` STRING,
> `boolean` BOOLEAN>))]
> ) WITH (
> 'connector' = 'filesystem',
> 'path' =
> 'file:///Users/davidradley/bashscripts/sqlscripts/json/test_avro.json',
> 'format' = 'avro'
> );
> {code}
>
> We get an ArrayIndexOutOfBoundsException_ exception thrown. The problem is
> [here]([https://github.com/apache/flink/blob/0261d498cb4538aefc4bb2f14ef08d0dec6db812/flink-table/flink-sql-parser/src/main/java/org/apache/flink/sql/parser/type/ExtendedSqlRowTypeNameSpec.java#L124)]
> where we address an element in the comments array, but the element does not
> exist. The proposed fix is to change this code to:
> {code:java}
> if (i < comments.size() && comments.get != null) {
> {code}
>
> There is one other place in the code base that has this if that I will also
> change.
>
> This does not occur at Flink 1.20.
>
>
> The stack trace is:
>
> {noformat}
>
> #18 1229.3 java.lang.ArrayIndexOutOfBoundsException
> #18 1229.3 at
> java.base/java.util.ImmutableCollections$ListN.get(ImmutableCollections.java:449)
> #18 1229.3 at
> org.apache.flink.sql.parser.type.ExtendedSqlRowTypeNameSpec.unparse(ExtendedSqlRowTypeNameSpec.java:124)
> #18 1229.3 at
> org.apache.calcite.sql.SqlDataTypeSpec.unparse(SqlDataTypeSpec.java:190)
> #18 1229.3 at
> org.apache.flink.sql.parser.type.ExtendedSqlRowTypeNameSpec.unparse(ExtendedSqlRowTypeNameSpec.java:120)
> #18 1229.3 at
> org.apache.calcite.sql.SqlDataTypeSpec.unparse(SqlDataTypeSpec.java:190)
> #18 1229.3 at
> org.apache.calcite.sql.fun.SqlCastFunction.unparse(SqlCastFunction.java:296)
> #18 1229.3 at
> org.apache.calcite.sql.SqlDialect.unparseCall(SqlDialect.java:466)
> #18 1229.3 at org.apache.calcite.sql.SqlCall.unparse(SqlCall.java:141)
> #18 1229.3 at
> org.apache.calcite.sql.fun.SqlMultisetValueConstructor.unparse(SqlMultisetValueConstructor.java:124)
> #18 1229.3 at
> org.apache.calcite.sql.SqlDialect.unparseCall(SqlDialect.java:466)
> #18 1229.3 at org.apache.calcite.sql.SqlCall.unparse(SqlCall.java:141)
> #18 1229.3 at org.apache.calcite.sql.SqlNode.toSqlString(SqlNode.java:156)
> #18 1229.3 at org.apache.calcite.sql.SqlNode.toSqlString(SqlNode.java:178)
> #18 1229.3 at org.apache.calcite.sql.SqlNode.toSqlString(SqlNode.java:187)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.SqlNodeConvertContext.toQuotedSqlString(SqlNodeConvertContext.java:110)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.table.SchemaBuilderUtil.toUnresolvedComputedColumn(SchemaBuilderUtil.java:213)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.table.MergeTableLikeUtil$SchemaBuilder.appendDerivedColumns(MergeTableLikeUtil.java:394)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.table.MergeTableLikeUtil.mergeTables(MergeTableLikeUtil.java:157)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.table.SqlCreateTableConverter$1.getMergedSchema(SqlCreateTableConverter.java:59)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.table.AbstractCreateTableConverter.getResolvedCatalogTable(AbstractCreateTableConverter.java:82)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.table.SqlCreateTableConverter.convertSqlNode(SqlCreateTableConverter.java:42)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.table.SqlCreateTableConverter.convertSqlNode(SqlCreateTableConverter.java:37)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.converters.SqlNodeConverters.convertSqlNode(SqlNodeConverters.java:127)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.SqlNodeToOperationConversion.convertValidatedSqlNode(SqlNodeToOperationConversion.java:231)
> #18 1229.3 at
> org.apache.flink.table.planner.operations.SqlNodeToOperationConversion.convert(SqlNodeToOperationConversion.java:221)
> #18 1229.3 at
> org.apache.flink.table.planner.delegation.ParserImpl.parse(ParserImpl.java:106)
> #18 1229.3 at
> org.apache.flink.table.api.internal.TableEnvironmentImpl.executeSql(TableEnvironmentImpl.java:937)
> #18 1229.3 at
> com.ibm.ei.streamproc.model.jobgraph.TemporaryTableCreator.createAndValidateTemporaryTable(TemporaryTableCreator.java:62)
> {noformat}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
