ExistingTableSchemaExpansionSupport (flink-cdc-parent 3.8-SNAPSHOT API)

All Superinterfaces:

Serializable

All Known Implementing Classes:

FlussMetaDataApplier, PaimonMetadataApplier
```
@Experimental
public interface ExistingTableSchemaExpansionSupport
extends Serializable
```
Connector-specific capabilities required to safely expand an existing target table schema.

Method Summary

All Methods Instance Methods Abstract Methods
Modifier and Type	Method	Description
`Optional<Schema>`	`getExistingTableSchema(TableId tableId)`	Returns the current schema of the target table, or `Optional.empty()` if the table does not exist.
`boolean`	`isColumnNameCaseSensitive()`	Returns whether target column names are matched case-sensitively.
`DataType`	`normalizeToTargetDataType(TableId tableId, String columnName, DataType pipelineDataType, Schema existingTargetSchema)`	Converts a pipeline type to the canonical CDC representation of the physical type that this applier will create in the existing target table.

- Method Detail
  - getExistingTableSchema
```
Optional<Schema> getExistingTableSchema(TableId tableId)
                                 throws SchemaEvolveException
```
    Returns the current schema of the target table, or Optional.empty() if the table does not exist. Target physical types must be converted to the same canonical CDC representation produced by normalizeToTargetDataType(TableId, String, DataType, Schema). The returned schema should include any table options that affect physical type mapping.
    
    Throws:
    
    SchemaEvolveException
  - normalizeToTargetDataType
```
DataType normalizeToTargetDataType(TableId tableId,
                                   String columnName,
                                   DataType pipelineDataType,
                                   Schema existingTargetSchema)
```
    Converts a pipeline type to the canonical CDC representation of the physical type that this applier will create in the existing target table. For example, if both CHAR(n) and VARCHAR(n) map to target VARCHAR(n), this method should return the same VARCHAR(n) representation for both input types. The existing target schema provides table options needed by sinks whose type mapping depends on table properties.
    This method must not query or modify the target table.
  - isColumnNameCaseSensitive
```
boolean isColumnNameCaseSensitive()
```
    Returns whether target column names are matched case-sensitively.