Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3829#discussion_r127452041
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/api/TableEnvironment.scala
---
@@ -488,19 +499,50 @@ abstract class TableEnvironment(val config:
TableConfig) {
* tEnv.sql(s"SELECT * FROM $table")
* }}}
*
- * @param query The SQL query to evaluate.
- * @return The result of the query as Table.
+ * @param sql The SQL string to evaluate.
+ * @return The result of the query as Table or null of the DML insert
operation.
*/
- def sql(query: String): Table = {
+ def sql(sql: String): Table = {
val planner = new FlinkPlannerImpl(getFrameworkConfig, getPlanner,
getTypeFactory)
// parse the sql query
- val parsed = planner.parse(query)
+ val parsed = planner.parse(sql)
// validate the sql query
val validated = planner.validate(parsed)
- // transform to a relational tree
- val relational = planner.rel(validated)
- new Table(this, LogicalRelNode(relational.rel))
+ // separate source translation if DML insert, considering sink
operation has no output but all
+ // RelNodes' translation should return a DataSet/DataStream, and
Calcite's TableModify
+ // defines output for DML(INSERT,DELETE,UPDATE) is a RowCount column
of BigInt type. This a
+ // work around since we have no conclusion for the output type
definition of sink operation
+ // (compare to traditional SQL insert).
+ if (parsed.isInstanceOf[SqlInsert]) {
+ val insert = parsed.asInstanceOf[SqlInsert]
+ // validate sink table
+ val targetName =
insert.getTargetTable.asInstanceOf[SqlIdentifier].names.get(0)
+ val targetTable = getTable(targetName)
+ if (null == targetTable ||
!targetTable.isInstanceOf[TableSinkTable[_]]) {
+ throw new TableException("SQL DML INSERT operation need a
registered TableSink Table!")
+ }
+ // validate unsupported partial insertion to sink table
+ val sinkTable = targetTable.asInstanceOf[TableSinkTable[_]]
+ if (null != insert.getTargetColumnList &&
insert.getTargetColumnList.size() !=
+ sinkTable.fieldTypes.length) {
--- End diff --
I think there is a mismatch between the `TableSink` interface and how it is
used as an output table here. An instance of `TableSink` has no predefined
columns but gets those when it is configured (`TableSink.configure()`) before
it emits a Table.
In the current state, a `TableSink` is configured to emit data according to
the query result, i.e, the schema of the emitted table depends on the query
result. However, for `INSERT INTO` the table schema is defined beforehand and
Calcite validates that the schema of the query is compliant with the schema of
the table to write to.
I think we should change the `TableSink` interface, but right now I don't
think we should use `TableSink` like that.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
