Package dev.frostlake.stream

Class StreamManager

java.lang.Object

dev.frostlake.stream.StreamManager

public class StreamManager
extends Object

Manages streams and tracks changes to tables

Constructor Summary

Constructors

Constructor	Description
StreamManager(Catalog catalog)

Method Summary

Modifier and Type	Method	Description
List<Stream>	capturingStreams(String qualifiedTableName)	The streams capturing qualifiedTableName, resolved ONCE for a bulk apply — commit used to re-walk every schema's streams per committed ROW.
void	onDatabaseDropped(String databaseName)	Forget the counters of every table in a dropped database (DROP DATABASE cascades).
void	onSchemaDropped(String databaseName, String schemaName)	Forget the counters of every table in a dropped schema (DROP SCHEMA cascades to its tables).
void	onTableDropped(String qualifiedTableName)	Forget a table's change-tracking row-id counter when the table is dropped, so rowIdCounters does not grow without bound as tables are created and dropped over the engine's lifetime.
void	seedInitialRows(Stream stream, String qualifiedTableName, List<Row> rows)	Seed a freshly created SHOW_INITIAL_ROWS stream with the source table's existing rows as INSERT records, so the stream's first read returns the current table contents (Snowflake semantics).
void	trackDelete(String qualifiedTableName, Row row)	Track a DELETE operation
void	trackDelete(List<Stream> capturing, String qualifiedTableName, Row row)	trackDelete(String, Row) over a PRE-RESOLVED capture list (bulk commit apply).
void	trackInsert(String qualifiedTableName, Row row)	Track an INSERT operation
void	trackInsert(List<Stream> capturing, String qualifiedTableName, Row row)	trackInsert(String, Row) over a PRE-RESOLVED capture list (bulk commit apply).
void	trackUpdate(String qualifiedTableName, Row oldRow, Row newRow)	Track an UPDATE operation
void	trackUpdate(List<Stream> capturing, String qualifiedTableName, Row oldRow, Row newRow)	trackUpdate(String, Row, Row) over a PRE-RESOLVED capture list (bulk commit apply).

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

StreamManager

public StreamManager(Catalog catalog)

Method Details

trackInsert

public void trackInsert(String qualifiedTableName,
 Row row)

Track an INSERT operation

capturingStreams

public List<Stream> capturingStreams(String qualifiedTableName)

The streams capturing qualifiedTableName, resolved ONCE for a bulk apply — commit used to re-walk every schema's streams per committed ROW. An empty answer lets the caller skip old-image copies and record construction entirely.

trackInsert

public void trackInsert(List<Stream> capturing,
 String qualifiedTableName,
 Row row)

trackInsert(String, Row) over a PRE-RESOLVED capture list (bulk commit apply).

trackUpdate

public void trackUpdate(List<Stream> capturing,
 String qualifiedTableName,
 Row oldRow,
 Row newRow)

trackUpdate(String, Row, Row) over a PRE-RESOLVED capture list (bulk commit apply).

trackDelete

public void trackDelete(List<Stream> capturing,
 String qualifiedTableName,
 Row row)

trackDelete(String, Row) over a PRE-RESOLVED capture list (bulk commit apply).

trackUpdate

public void trackUpdate(String qualifiedTableName,
 Row oldRow,
 Row newRow)

Track an UPDATE operation

trackDelete

public void trackDelete(String qualifiedTableName,
 Row row)

Track a DELETE operation

seedInitialRows

public void seedInitialRows(Stream stream,
 String qualifiedTableName,
 List<Row> rows)

Seed a freshly created SHOW_INITIAL_ROWS stream with the source table's existing rows as INSERT records, so the stream's first read returns the current table contents (Snowflake semantics). Subsequent reads after consumption then return only new changes as usual.

onTableDropped

public void onTableDropped(String qualifiedTableName)

Forget a table's change-tracking row-id counter when the table is dropped, so rowIdCounters does not grow without bound as tables are created and dropped over the engine's lifetime. Keyed by the fully-qualified name, normalized to upper case to match the write path (getNextRowId(java.lang.String)).

onSchemaDropped

public void onSchemaDropped(String databaseName,
 String schemaName)

Forget the counters of every table in a dropped schema (DROP SCHEMA cascades to its tables).

onDatabaseDropped

public void onDatabaseDropped(String databaseName)

Forget the counters of every table in a dropped database (DROP DATABASE cascades).