序
本文主要研究一下debezium的RecordMakers
RecordMakers
debezium-v1.1.1.Final/debezium-connector-mysql/src/main/java/io/debezium/connector/mysql/RecordMakers.java
public class RecordMakers {
private final Logger logger = LoggerFactory.getLogger(getClass());
private final MySqlSchema schema;
private final SourceInfo source;
private final TopicSelector topicSelector;
private final boolean emitTombstoneOnDelete;
private final Map convertersByTableNumber = new HashMap<>();
private final Map tableNumbersByTableId = new HashMap<>();
private final Map tableIdsByTableNumber = new HashMap<>();
private final Schema schemaChangeKeySchema;
private final Schema schemaChangeValueSchema;
private final SchemaNameAdjuster schemaNameAdjuster = SchemaNameAdjuster.create(logger);
private final Map restartOffset;
//......
public RecordMakers(MySqlSchema schema, SourceInfo source, TopicSelector topicSelector,
boolean emitTombstoneOnDelete, Map restartOffset) {
this.schema = schema;
this.source = source;
this.topicSelector = topicSelector;
this.emitTombstoneOnDelete = emitTombstoneOnDelete;
this.restartOffset = restartOffset;
this.schemaChangeKeySchema = SchemaBuilder.struct()
.name(schemaNameAdjuster.adjust("io.debezium.connector.mysql.SchemaChangeKey"))
.field(Fields.DATABASE_NAME, Schema.STRING_SCHEMA)
.build();
this.schemaChangeValueSchema = SchemaBuilder.struct()
.name(schemaNameAdjuster.adjust("io.debezium.connector.mysql.SchemaChangeValue"))
.field(Fields.SOURCE, source.schema())
.field(Fields.DATABASE_NAME, Schema.STRING_SCHEMA)
.field(Fields.DDL_STATEMENTS, Schema.STRING_SCHEMA)
.build();
}
public RecordsForTable forTable(TableId tableId, BitSet includedColumns, BlockingConsumer consumer) {
Long tableNumber = tableNumbersByTableId.get(tableId);
return tableNumber != null ? forTable(tableNumber, includedColumns, consumer) : null;
}
//......
} - RecordMakers提供了forTable方法，用于创建RecordsForTable
RecordsForTable
debezium-v1.1.1.Final/debezium-connector-mysql/src/main/java/io/debezium/connector/mysql/RecordMakers.java
public final class RecordsForTable {
private final BitSet includedColumns;
private final Converter converter;
private final BlockingConsumer consumer;
protected RecordsForTable(Converter converter, BitSet includedColumns, BlockingConsumer consumer) {
this.converter = converter;
this.includedColumns = includedColumns;
this.consumer = consumer;
}
public int read(Object[] row, Instant ts) throws InterruptedException {
return read(row, ts, 0, 1);
}
public int read(Object[] row, Instant ts, int rowNumber, int numberOfRows) throws InterruptedException {
return converter.read(source, row, rowNumber, numberOfRows, includedColumns, ts, consumer);
}
public int create(Object[] row, Instant ts) throws InterruptedException {
return create(row, ts, 0, 1);
}
public int create(Object[] row, Instant ts, int rowNumber, int numberOfRows) throws InterruptedException {
return converter.insert(source, row, rowNumber, numberOfRows, includedColumns, ts, consumer);
}
public int update(Object[] before, Object[] after, Instant ts) throws InterruptedException {
return update(before, after, ts, 0, 1);
}
public int update(Object[] before, Object[] after, Instant ts, int rowNumber, int numberOfRows) throws InterruptedException {
return converter.update(source, before, after, rowNumber, numberOfRows, includedColumns, ts, consumer);
}
public int delete(Object[] row, Instant ts) throws InterruptedException {
return delete(row, ts, 0, 1);
}
public int delete(Object[] row, Instant ts, int rowNumber, int numberOfRows) throws InterruptedException {
return converter.delete(source, row, rowNumber, numberOfRows, includedColumns, ts, consumer);
}
} - RecordsForTable提供了read、create、update、delete方法，它们都委托给converter的对应方法
Converter
debezium-v1.1.1.Final/debezium-connector-mysql/src/main/java/io/debezium/connector/mysql/RecordMakers.java
protected static interface Converter {
int read(SourceInfo source, Object[] row, int rowNumber, int numberOfRows, BitSet includedColumns, Instant ts,
BlockingConsumer consumer)
throws InterruptedException;
int insert(SourceInfo source, Object[] row, int rowNumber, int numberOfRows, BitSet includedColumns, Instant ts,
BlockingConsumer consumer)
throws InterruptedException;
int update(SourceInfo source, Object[] before, Object[] after, int rowNumber, int numberOfRows, BitSet includedColumns, Instant ts,
BlockingConsumer consumer)
throws InterruptedException;
int delete(SourceInfo source, Object[] row, int rowNumber, int numberOfRows, BitSet includedColumns, Instant ts,
BlockingConsumer consumer)
throws InterruptedException;
} - Converter接口定义了read、insert、update、delete方法
RecordMakers.Converter
debezium-v1.1.1.Final/debezium-connector-mysql/src/main/java/io/debezium/connector/mysql/RecordMakers.java
Converter converter = new Converter() {
@Override
public int read(SourceInfo source, Object[] row, int rowNumber, int numberOfRows, BitSet includedColumns, Instant ts,
BlockingConsumer consumer)
throws InterruptedException {
Object key = tableSchema.keyFromColumnData(row);
Struct value = tableSchema.valueFromColumnData(row);
if (value != null || key != null) {
Schema keySchema = tableSchema.keySchema();
Map partition = source.partition();
Map offset = source.offsetForRow(rowNumber, numberOfRows);
source.tableEvent(id);
Struct origin = source.struct();
SourceRecord record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum,
keySchema, key, envelope.schema(), envelope.read(value, origin, ts));
consumer.accept(record);
return 1;
}
return 0;
}
@Override
public int insert(SourceInfo source, Object[] row, int rowNumber, int numberOfRows, BitSet includedColumns, Instant ts,
BlockingConsumer consumer)
throws InterruptedException {
validateColumnCount(tableSchema, row);
Object key = tableSchema.keyFromColumnData(row);
Struct value = tableSchema.valueFromColumnData(row);
if (value != null || key != null) {
Schema keySchema = tableSchema.keySchema();
Map partition = source.partition();
Map offset = source.offsetForRow(rowNumber, numberOfRows);
source.tableEvent(id);
Struct origin = source.struct();
SourceRecord record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum,
keySchema, key, envelope.schema(), envelope.create(value, origin, ts));
consumer.accept(record);
return 1;
}
return 0;
}
@Override
public int update(SourceInfo source, Object[] before, Object[] after, int rowNumber, int numberOfRows, BitSet includedColumns,
Instant ts,
BlockingConsumer consumer)
throws InterruptedException {
int count = 0;
validateColumnCount(tableSchema, after);
Object key = tableSchema.keyFromColumnData(after);
Struct valueAfter = tableSchema.valueFromColumnData(after);
if (valueAfter != null || key != null) {
Object oldKey = tableSchema.keyFromColumnData(before);
Struct valueBefore = tableSchema.valueFromColumnData(before);
Schema keySchema = tableSchema.keySchema();
Map partition = source.partition();
Map offset = source.offsetForRow(rowNumber, numberOfRows);
source.tableEvent(id);
Struct origin = source.struct();
if (key != null && !Objects.equals(key, oldKey)) {
// The key has changed, so we need to deal with both the new key and old key.
// Consumers may push the events into a system that won't allow both records to exist at the same time,
// so we first want to send the delete event for the old key...
SourceRecord record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum,
keySchema, oldKey, envelope.schema(), envelope.delete(valueBefore, origin, ts));
consumer.accept(record);
++count;
if (emitTombstoneOnDelete) {
// Next send a tombstone event for the old key ...
record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum, keySchema, oldKey, null, null);
consumer.accept(record);
++count;
}
// And finally send the create event ...
record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum,
keySchema, key, envelope.schema(), envelope.create(valueAfter, origin, ts));
consumer.accept(record);
++count;
}
else {
// The key has not changed, so a simple update is fine ...
SourceRecord record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum,
keySchema, key, envelope.schema(), envelope.update(valueBefore, valueAfter, origin, ts));
consumer.accept(record);
++count;
}
}
return count;
}
@Override
public int delete(SourceInfo source, Object[] row, int rowNumber, int numberOfRows, BitSet includedColumns, Instant ts,
BlockingConsumer consumer)
throws InterruptedException {
int count = 0;
validateColumnCount(tableSchema, row);
Object key = tableSchema.keyFromColumnData(row);
Struct value = tableSchema.valueFromColumnData(row);
if (value != null || key != null) {
Schema keySchema = tableSchema.keySchema();
Map partition = source.partition();
Map offset = source.offsetForRow(rowNumber, numberOfRows);
source.tableEvent(id);
Struct origin = source.struct();
// Send a delete message ...
SourceRecord record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum,
keySchema, key, envelope.schema(), envelope.delete(value, origin, ts));
consumer.accept(record);
++count;
// And send a tombstone ...
if (emitTombstoneOnDelete) {
record = new SourceRecord(partition, getSourceRecordOffset(offset), topicName, partitionNum,
keySchema, key, null, null);
consumer.accept(record);
++count;
}
}
return count;
}
@Override
public String toString() {
return "RecordMaker.Converter(" + id + ")";
}
private void validateColumnCount(TableSchema tableSchema, Object[] row) {
final int expectedColumnsCount = schema.tableFor(tableSchema.id()).columns().size();
if (expectedColumnsCount != row.length) {
logger.error("Invalid number of columns, expected '{}' arrived '{}'", expectedColumnsCount, row.length);
throw new ConnectException(
"The binlog event does not contain expected number of columns; the internal schema representation is probably out of sync with the real database schema, or the binlog contains events recorded with binlog_row_image other than FULL or the table in question is an NDB table");
}
}
}; - RecordMakers创建了一个匿名Converter实现类，其实现方法基本是创建kafka connect的SourceRecord，然后执行consumer.accept(record)
小结
RecordMakers提供了forTable方法，用于创建RecordsForTable；RecordsForTable提供了read、create、update、delete方法，它们都委托给converter的对应方法