[FLINK-36796][pipeline-connector][oracle]add oracle pipeline connector. by linjianchang · Pull Request #3995 · apache/flink-cdc

linjianchang · 2025-04-18T09:07:19Z

add oracle pipeline connector.

joyCurry30

Thank you for your contribution. I just left some comment.

joyCurry30 · 2025-04-19T00:48:31Z

+        <dependency>
+            <groupId>io.debezium</groupId>
+            <artifactId>debezium-core</artifactId>
+            <version>1.9.8.Final</version>


Use ${debezium.version}

And the scope should be "provide".

Have been modified

joyCurry30 · 2025-04-19T00:51:27Z

+        <dependency>
+            <groupId>com.ververica</groupId>
+            <artifactId>flink-cdc-source-e2e-tests</artifactId>
+            <version>cty-3.0-2.2-SNAPSHOT</version>


Why you dependent on "flink-cdc-source-e2e-tests"?

Have been removed

joyCurry30 · 2025-04-19T00:52:08Z

+        <dependency>
+            <groupId>org.apache.flink</groupId>
+            <artifactId>flink-connector-test-util</artifactId>
+            <version>3.4-SNAPSHOT</version>


Please use "${project.version}".

Have been modified

joyCurry30 · 2025-04-19T00:56:02Z

+                                    <include>io.debezium:debezium-core</include>
+                                    <include>io.debezium:debezium-ddl-parser</include>
+                                    <include>io.debezium:debezium-connector-oracle</include>
+                                    <include>io.debezium:debezium-connector-mysql</include>


Could you explain the rationale behind having a dependency on the MySQL CDC connector within the Oracle CDC connector implementation? I'd like to better understand how these components interact in this context.

Have been removed

joyCurry30 · 2025-04-19T00:56:24Z

+                                    <include>io.debezium:debezium-connector-oracle</include>
+                                    <include>io.debezium:debezium-connector-mysql</include>
+                                    <include>com.ververica:flink-connector-debezium</include>
+                                    <include>com.ververica:flink-connector-mysql-cdc</include>


Same as above.

Have been removed

joyCurry30 · 2025-04-19T00:56:49Z

+                                    <include>org.antlr:antlr4-runtime</include>
+                                    <include>org.apache.kafka:*</include>
+                                    <include>mysql:mysql-connector-java</include>
+                                    <include>com.zendesk:mysql-binlog-connector-java</include>


Same as above.

Have been removed

joyCurry30 · 2025-04-19T01:08:37Z

+    public static final ConfigOption<String> JDBC_URL =
+            ConfigOptions.key("jdbc.url")
+                    .stringType()
+                    .noDefaultValue()
+                    .withDescription("The jdbc url.");


Could we clarify the relationship between the JDBC URL and the individual hostname/port parameters?

If we already have a “jdbc.url” configuration field, is there still value in maintaining separate “hostname and “port” parameters? Should these be mutually exclusive?

When using “hostname” and "port" configuration, is there a parameter to explicitly specify the driver type (Thin vs OCI)? This would be crucial for constructing the correct JDBC connection string format.

In The original oracle module [flink-connector-oracle-cdc]OracleJdbcUrlUtils#getConnectionUrlWithSid method，url，hostname and port is not exclusive，when url config is null，url = "jdbc:oracle:thin:@" + hostname + ":" + port + ":" + dbname;，but 19c url is must connect through the service name，like url = "jdbc:oracle:thin:@" + hostname + ":" + port + "/" + dbname; So you can configure the URL to adapt to oracle 19c

joyCurry30 · 2025-04-19T01:14:43Z

+        if (isAddMeta) {
+            map.put(OracleDataSourceOptions.HOSTNAME.key(), hostname);
+            map.put(OracleDataSourceOptions.PORT.key(), port);
+        }


Could you clarify the design intent behind treating hostname and port as common metadata fields?

Since we don't currently support multi-source CDC synchronization, are these fields intended for future extensibility?

Have been removed

linjianchang · 2025-04-23T01:19:18Z

@joyCurry30 Already modified，please review again ,thanks!

joyCurry30

Hi, thank you for your contribution. I left some comments for the doc.

joyCurry30 · 2025-04-27T01:58:30Z

+
+```yaml
+source:
+  type: mysql


“type” should be "oracle".

have modified

joyCurry30 · 2025-04-27T01:59:19Z

+- `initial` （默认）：在第一次启动时对受监视的数据库表执行初始快照，并继续读取最新的 binlog。
+- `latest-offset`：首次启动时，从不对受监视的数据库表执行快照， 连接器仅从 binlog 的结尾处开始读取，这意味着连接器只能读取在连接器启动之后的数据更改。


Does oracle cdc use binlog?

have modified

joyCurry30 · 2025-04-27T02:01:52Z

+    <tr>
+      <td>
+       XMLTYPE
+        </td>
+      <td>VARCHAR(n)</td>
+      <td></td>
+    </tr>
+    <tr>
+      <td>
+        VARCHAR(n)<br>
+        VARCHAR2(n)<br>
+        NVARCHAR2(n)<br>
+        NCHAR(n)<br>
+        CHAR(n)<br>
+      </td>
+      <td>VARCHAR(n)</td>
+      <td></td>
+    </tr>


These two elements should be merged.

have modified

joyCurry30 · 2025-04-27T02:09:56Z

+
+```yaml
+source:
+  type: mysql


Same as above.

have modified

joyCurry30 · 2025-04-27T02:09:56Z

+- `initial` (default): Performs an initial snapshot on the monitored database tables upon first startup, and continue to read the latest binlog.
+- `latest-offset`: Never to perform snapshot on the monitored database tables upon first startup, just read from
+


Same as above.

have modified

joyCurry30 · 2025-04-27T02:10:17Z

+## 数据类型映射
+


Use English, please.

have modified

joyCurry30 · 2025-07-03T02:21:48Z

+                                .hostname(
+                                        config.getOptional(OracleDataSourceOptions.HOSTNAME).get())
+                                .port(config.getOptional(OracleDataSourceOptions.PORT).get())
+                                .databaseList(
+                                        config.getOptional(OracleDataSourceOptions.DATABASE)
+                                                .get()) // monitor oracledatabase
+                                .tableList(
+                                        config.getOptional(OracleDataSourceOptions.TABLES)
+                                                .get()) // monitor productstable
+                                .username(
+                                        config.getOptional(OracleDataSourceOptions.USERNAME).get())
+                                .password(
+                                        config.getOptional(OracleDataSourceOptions.PASSWORD).get())
+                                .includeSchemaChanges(true);


Use config.get(ConfigOption option).

have modified

joyCurry30 · 2025-07-03T02:24:57Z

+        options.add(OracleDataSourceOptions.SCHEMALIST);
+        options.add(OracleDataSourceOptions.DATABASE);
+        options.add(OracleDataSourceOptions.TABLES);
+        options.add(METADATA_LIST);


Use OracleDataSourceOptions.METADATA_LIST.

have modified

joyCurry30 · 2025-07-03T02:26:40Z

+    private static final String SCAN_STARTUP_MODE_VALUE_INITIAL = "initial";
+    private static final String SCAN_STARTUP_MODE_VALUE_EARLIEST = "earliest-offset";
+    private static final String SCAN_STARTUP_MODE_VALUE_LATEST = "latest-offset";
+    private static final String SCAN_STARTUP_MODE_VALUE_SPECIFIC_OFFSET = "specific-offset";
+    private static final String SCAN_STARTUP_MODE_VALUE_TIMESTAMP = "timestamp";


Use enum replace switch-case.

reference mysql pipeline，it like this

joyCurry30 · 2025-07-03T02:30:38Z

+                    builder.hostname(config.getOptional(OracleDataSourceOptions.HOSTNAME).get())
+                            .port(config.getOptional(OracleDataSourceOptions.PORT).get())
+                            .database(
+                                    config.getOptional(OracleDataSourceOptions.DATABASE)
+                                            .get()) // monitor  database
+                            .schemaList(
+                                    config.getOptional(OracleDataSourceOptions.SCHEMALIST)
+                                            .get()) // monitor  schema
+                            .tableList(capturedTables) // monitor
+                            // EMP table
+                            .username(config.getOptional(OracleDataSourceOptions.USERNAME).get())
+                            .password(config.getOptional(OracleDataSourceOptions.PASSWORD).get())


Use config.get(ConfigOption option)

have modified

aiwenmo · 2025-07-09T01:54:18Z

+                                    <shadedPattern>
+                                        com.ververica.cdc.connectors.shaded.org.apache.kafka
+                                    </shadedPattern>
+                                </relocation>
+                                <relocation>
+                                    <pattern>org.antlr</pattern>
+                                    <shadedPattern>
+                                        com.ververica.cdc.connectors.shaded.org.antlr
+                                    </shadedPattern>
+                                </relocation>
+                                <relocation>
+                                    <pattern>com.fasterxml</pattern>
+                                    <shadedPattern>
+                                        com.ververica.cdc.connectors.shaded.com.fasterxml
+                                    </shadedPattern>
+                                </relocation>
+                                <relocation>
+                                    <pattern>com.google</pattern>
+                                    <shadedPattern>
+                                        com.ververica.cdc.connectors.shaded.com.google
+                                    </shadedPattern>
+                                </relocation>
+                                <relocation>
+                                    <pattern>com.esri.geometry</pattern>
+                                    <shadedPattern>com.ververica.cdc.connectors.shaded.com.esri.geometry</shadedPattern>
+                                </relocation>
+                                <relocation>
+                                    <pattern>com.zaxxer</pattern>
+                                    <shadedPattern>
+                                        com.ververica.cdc.connectors.shaded.com.zaxxer
+                                    </shadedPattern>


Why relocate com.ververica.cdc.connectors?
Please use the correct package path.

have modified

aiwenmo · 2025-07-09T01:55:58Z

+package com.apache.flink.cdc.connectors.oracle.dto;
+


Please use org.apache.

have modified

aiwenmo · 2025-07-09T01:56:59Z

+import com.apache.flink.cdc.connectors.oracle.source.OracleDataSource;
+import com.apache.flink.cdc.connectors.oracle.source.OracleDataSourceOptions;
+import com.apache.flink.cdc.connectors.oracle.utils.OracleSchemaUtils;


Please use org.apache.

have modified

aiwenmo · 2025-07-09T01:57:26Z

+import java.util.Set;
+import java.util.stream.Collectors;
+
+import static com.apache.flink.cdc.connectors.oracle.source.OracleDataSourceOptions.METADATA_LIST;


Please use org.apache.

have modified

aiwenmo · 2025-07-09T02:00:07Z

+        }
+        throw new IllegalArgumentException(
+                String.format(
+                        "[%s] cannot be found in mysql metadata.",


oracle metadata

thanks for review！have modified！

aiwenmo · 2025-10-12T05:16:20Z

+        dbzProperties.setProperty(
+                "snapshot.locking.mode", config.get(OracleDataSourceOptions.SNAPSHOT_LOCKING_MODE));
+        dbzProperties.setProperty(
+                "snapshot.locking.mode", config.get(OracleDataSourceOptions.SNAPSHOT_LOCKING_MODE));


This is a duplicate code snippet.

have remove the duplicate code

aiwenmo · 2025-10-12T05:20:51Z

+    @Override
+    public EventSourceProvider getEventSourceProvider() {
+        String url = config.get(OracleDataSourceOptions.JDBC_URL);


How should null parameters be handled?

url could be null，and hostname，port，database，username，password must be not null，have add checknotnull

aiwenmo · 2025-10-12T05:23:37Z

+        switch (modeString.toLowerCase()) {
+            case SCAN_STARTUP_MODE_VALUE_INITIAL:
+                return StartupOptions.initial();
+
+            case SCAN_STARTUP_MODE_VALUE_LATEST:
+                return StartupOptions.latest();
+


How should other situations of StartupOptions be handled?

We have added the remaining supported StartupOptions

aiwenmo · 2025-10-12T05:26:11Z

+    public static final ConfigOption<String> SERVER_TIME_ZONE =
+            ConfigOptions.key("server-time-zone")
+                    .stringType()
+                    .noDefaultValue()
+                    .withDescription(
+                            "The session time zone in database server. If not set, then "
+                                    + "ZoneId.systemDefault() is used to determine the server time zone.");
+
+    public static final ConfigOption<String> SERVER_ID =
+            ConfigOptions.key("server-id")
+                    .stringType()


Does Oracle CDC have these configurations?

have remove SERVER_ID，and SERVER_TIME_ZONE is need

aiwenmo · 2025-10-12T05:28:52Z

+    @Override
+    protected Object convertToString(Object dbzObj, Schema schema) {
+        // the Geometry datatype in oracle will be converted to
+        // a String with Json format
+        if (Point.LOGICAL_NAME.equals(schema.name())


I'm not sure if Oracle also needs to be handled in this way.

Oracle database provides powerful geometric data type support through the Spatial module, which is mainly based on the SDO_GEOMETRY object type.

aiwenmo · 2025-10-12T05:32:47Z

Hi. Thanks for your contribution. @linjianchang

I've found that some code might be useless in Oracle CDC. Please simplify the code carefully to make it as concise as possible.

Mrart · 2025-12-15T01:51:05Z

and we need implement e2e OracleE2eITCase like PostgresE2eITCase.

Aleady add OracleE2eITCase

LGTM. cc @LYanquan

lvyanquan

Thanks @linjianchang's contribution, left some comments.

lvyanquan · 2025-12-25T09:42:54Z

@@ -0,0 +1,15 @@
+# Copyright 2023 Ververica Inc.


We should use ASF License.

We should use ASF License.

modified

lvyanquan · 2025-12-25T09:46:55Z

+
+# Set root logger level to OFF to not flood build logs
+# set manually to INFO for debugging purposes
+rootLogger.level=INFO


It's better to set this to ERROR to reduce log in CI machine.

It's better to set this to ERROR to reduce log in CI machine.

modified

lvyanquan · 2025-12-25T09:48:17Z

@@ -0,0 +1,266 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!--
+  ~ Copyright 2023 Ververica Inc.


Should use ASF license.

Should use ASF license.

modified

lvyanquan · 2025-12-25T09:56:17Z

@@ -0,0 +1,26 @@
+################################################################################
+#  Copyright 2023 Ververica Inc.


Ditto. Maybe you could modify all files at once.

Ditto. Maybe you could modify all files at once.

modified

lvyanquan · 2025-12-25T11:11:30Z

+                    .durationType()
+                    .defaultValue(Duration.ofSeconds(30))
+                    .withDescription(
+                            "Optional interval of sending heartbeat event for tracing the latest available binlog offsets");


binlog?

modified

lvyanquan · 2025-12-25T11:20:12Z

+            ConfigOptions.key("debezium.database.connection.adapter")
+                    .stringType()
+                    .defaultValue("logminer")
+                    .withDescription("Database connection adapter.");


Should introduce all available values.

Should introduce all available values.

modified

lvyanquan · 2025-12-25T11:26:52Z

+
+/** IT tests for {@link OracleDataSource}. */
+@TestMethodOrder(MethodOrderer.OrderAnnotation.class)
+public class OraclePipelineITCase extends OracleSourceTestBase {


It's necessary to add some tests to cover the scene of restoring from savepoint with different start up mode.

It's necessary to add some tests to cover the scene of restoring from savepoint with different start up mode.

added

lvyanquan · 2025-12-25T11:29:31Z

            String str = (String) dbzObj;
            // TIMESTAMP_LTZ type is encoded in string type
-            Instant instant = Instant.parse(str);
+            Instant instant = ZonedTimestamp.FORMATTER.parse(str, Instant::from);


Could you add some explanation about why this change is necessary?

Could you add some explanation about why this change is necessary?

The parsing of the time zone type for Oracle failed

All sources in the pipeline have been tested with fulltype's ut and are normal. This change has not affected other connectors.

lvyanquan · 2025-12-25T11:32:23Z

+                                + "  port: %d\n"
+                                + "  username: %s\n"
+                                + "  password: %s\n"
+                                + "  tables: %s.PRODUCTS\n"


There is only one table here, you can add a few more tables.

There is only one table here, you can add a few more tables.

modified

lvyanquan · 2025-12-25T11:41:44Z

                return DataTypes.STRING();
            }
+            if (scale < 0 || scale > 36) {
+                return DataTypes.STRING();


Please explain why this change is necessary.

Please explain why this change is necessary.

"Does Oracle's 'nember' type have a Scala value?" It is in the range of 0-36 in decimal, so it is not in the decimal range Convert to string
All sources in the pipeline have been tested with fulltype's ut and are normal. This change has not affected other connectors.

lvyanquan · 2025-12-26T01:51:09Z

+        OracleDialect oracleDialect = new OracleDialect();
+        OracleEventDeserializer deserializer =
+                new OracleEventDeserializer(
+                        DebeziumChangelogMode.ALL,
+                        config.get(OracleDataSourceOptions.SCHEMA_CHANGE_ENABLED),
+                        readableMetadataList);
+
+        RedoLogOffsetFactory offsetFactory = new RedoLogOffsetFactory();
+        OracleTableSourceReader oracleChangeEventSource =
+                new OracleTableSourceReader(
+                        configFactory, deserializer, offsetFactory, oracleDialect);


Can be updated to the following code to avoid Raw use of parameterized class.

OracleEventDeserializer<Event> deserializer = new OracleEventDeserializer<>( DebeziumChangelogMode.ALL, config.get(OracleDataSourceOptions.SCHEMA_CHANGE_ENABLED), readableMetadataList); RedoLogOffsetFactory offsetFactory = new RedoLogOffsetFactory(); OracleTableSourceReader<Event> oracleChangeEventSource = new OracleTableSourceReader<>( configFactory, deserializer, offsetFactory, oracleDialect);

Can be updated to the following code to avoid Raw use of parameterized class.

OracleEventDeserializer<Event> deserializer = new OracleEventDeserializer<>( DebeziumChangelogMode.ALL, config.get(OracleDataSourceOptions.SCHEMA_CHANGE_ENABLED), readableMetadataList); RedoLogOffsetFactory offsetFactory = new RedoLogOffsetFactory(); OracleTableSourceReader<Event> oracleChangeEventSource = new OracleTableSourceReader<>( configFactory, deserializer, offsetFactory, oracleDialect);

Modified

lvyanquan · 2025-12-26T02:03:58Z

+import static org.apache.flink.cdc.connectors.base.utils.SourceRecordUtils.isSchemaChangeEvent;
+
+/** The {@link RecordEmitter} implementation for Oracle pipeline connector. */
+public class OraclePipelineRecordEmitter<T> extends IncrementalSourceRecordEmitter<T> {


We can clarify that the generic type T here is Event.

We can clarify that the generic type T here is Event.

Modified

lvyanquan · 2025-12-26T02:04:19Z

+ * and Watermark Signal Algorithm which supports parallel reading snapshot of table and then
+ * continue to capture data change by streaming reading.
+ */
+public class OracleTableSourceReader<T> extends OracleSourceBuilder.OracleIncrementalSource<T> {


We can clarify that the generic type T here is Event.

We can clarify that the generic type T here is Event.

Modified

lvyanquan · 2025-12-26T02:14:11Z

+            if (isDataChangeRecord || isSchemaChangeEvent(element)) {
+                TableId debeziumTableId = getTableId(element);
+                TableId tableId =
+                        TableId.parse(debeziumTableId.schema() + "." + debeziumTableId.table());


Can we use debeziumTableId directly here?

Can we use debeziumTableId directly here?

Modified

lvyanquan · 2025-12-26T02:14:55Z

+                    sendCreateTableEvent(jdbc, tableId, output);
+                }
+                alreadySendCreateTableTables.add(
+                        TableId.parse(tableId.schema() + "." + tableId.table()));


alreadySendCreateTableTables.add(new TableId(tableId.schema(), null, tableId.table()));

alreadySendCreateTableTables.add(new TableId(tableId.schema(), null, tableId.table()));

Modified

lvyanquan · 2025-12-26T02:18:48Z

+
+    private static final Logger LOGGER =
+            LoggerFactory.getLogger(
+                    io.debezium.connector.oracle.antlr.listener.AlterTableParserListener.class);


LoggerFactory.getLogger(OracleAlterTableParserListener.class);

LoggerFactory.getLogger(OracleAlterTableParserListener.class);

Modified

lvyanquan · 2025-12-26T02:28:40Z

+                    .booleanType()
+                    .defaultValue(true)
+                    .withDescription(
+                            "Whether send schema change events, by default is true. If set to false, the schema changes will not be sent.");


What would happen if we alter table structure but setting schema-change.enabled to false? Should we always set this parameter to true?

What would happen if we alter table structure but setting schema-change.enabled to false? Should we always set this parameter to true?
always set this parameter to true

lvyanquan · 2025-12-26T02:32:59Z

+/**
+ * Test utility for creating converter, formatter and deserializer of a table in the test database.
+ */
+public class TestTable {


This class seems to be useless.

This class seems to be useless.

already removed

lvyanquan · 2025-12-26T02:33:14Z

+import static org.apache.flink.util.Preconditions.checkState;
+
+/** Oracle test utilities. */
+public class OracleTestUtils {


This class seems to be useless.

This class seems to be useless.

already removed

lvyanquan · 2025-12-26T02:33:37Z

+import java.util.stream.Collectors;
+
+/** Formatter that formats the {@link SourceRecord} to String. */
+public class RecordsFormatter {


This class seems to be useless.

This class seems to be useless.

already removed

lvyanquan · 2025-12-26T02:48:57Z

+            throw new IllegalArgumentException(
+                    "Cannot find any table by the option 'tables' = " + tables);
+        }
+        configFactory.tableList(capturedTables);


In this configuration mode, we cannot handle tables newly created during the incremental phase. Perhaps you can refer to the implementation of MySQL.

In this configuration mode, we cannot handle tables newly created during the incremental phase. Perhaps you can refer to the implementation of MySQL.

Remove it temporarily. We won't add this feature in this version

zml1206 · 2026-01-29T08:31:19Z

+
+    @Override
+    public void enterModify_col_properties(PlSqlParser.Modify_col_propertiesContext ctx) {
+        resolveColumnDataType(ctx);


If dataType is not specified, will changing the column attribute to nullable have no effect?

In this case, could we add a drop column event? Drop column seems simply makes the column nullable.

If dataType is not specified, will changing the column attribute to nullable have no effect?

Why is there a situation where no data type is specified? If no new data type is specified in the MODIFY clause, Oracle will report an error

ALTER TABLE t MODIFY a NULL;
Is this syntax not supported?

https://docs.oracle.com/en/database/oracle/oracle-database/19/sqlrf/img_text/modify_col_properties.html
dataType is optional.

lvyanquan · 2026-02-03T08:08:07Z

Please update https://github.com/apache/flink-cdc/blob/master/.github/workflows/modules.py to include test for this newly added module.

linjianchang · 2026-02-03T09:41:30Z

Please update https://github.com/apache/flink-cdc/blob/master/.github/workflows/modules.py to include test for this newly added module.

Have updated the file in master-36796

… com.fasterxml.jackson.

lvyanquan

+1.

paulo-t · 2026-02-11T07:49:52Z

+    public static List<TableId> listTables(
+            OracleSourceConfig sourceConfig, @Nullable String dbName) {
+        try (JdbcConnection jdbc = createOracleConnection(sourceConfig)) {
+            List<String> databases =


CDB + PDB mode can not find PDB databases here, should change session to PDB through setSessionToPdb(${pdbName}) before listDatabase?

…r. (apache#3995)

sumitnekar · 2026-04-04T18:15:33Z

+        Set<TableId> tableIdSet = new HashSet<>();
+        String queryTablesSql =
+                "SELECT OWNER ,TABLE_NAME,TABLESPACE_NAME FROM ALL_TABLES \n"
+                        + "WHERE TABLESPACE_NAME IS NOT NULL AND TABLESPACE_NAME NOT IN ('SYSTEM','SYSAUX') "


Oracle returns table space as null for partitioned tables. When I tested with a partitioned table, this table was never discovered and initial snapshot never started. So it was throwing generic error - “Db history topic or its content is fully or partially missing. Please check ur database history topic configuration”.

But it worked fine for me with
SourceFunction-based DataStream API. It never worked for me with Incremental Snapshot based DataStream API.

@linjianchang @lvyanquan Let me know If I am missing something.

github-actions Bot added base oracle-cdc-connector debezium labels Apr 18, 2025

joyCurry30 reviewed Apr 19, 2025

View reviewed changes

linjianchang force-pushed the master-36796 branch from 820b720 to 5015bbb Compare April 21, 2025 06:27

github-actions Bot added the docs Improvements or additions to documentation label Apr 21, 2025

joyCurry30 reviewed Apr 27, 2025

View reviewed changes

linjianchang force-pushed the master-36796 branch from 5015bbb to f284d27 Compare April 27, 2025 02:55

linjianchang force-pushed the master-36796 branch from f284d27 to f6b25cc Compare May 9, 2025 09:26

linjianchang force-pushed the master-36796 branch from f6b25cc to 4f73a6f Compare June 12, 2025 10:38

joyCurry30 reviewed Jul 3, 2025

View reviewed changes

linjianchang force-pushed the master-36796 branch from 4f73a6f to 3479f43 Compare July 4, 2025 03:35

aiwenmo suggested changes Jul 9, 2025

View reviewed changes

linjianchang force-pushed the master-36796 branch from 13c6586 to ba183e3 Compare July 10, 2025 06:07

linjianchang force-pushed the master-36796 branch 2 times, most recently from 0c77212 to 0237401 Compare August 5, 2025 10:39

linjianchang requested review from aiwenmo and joyCurry30 October 10, 2025 08:26

aiwenmo reviewed Oct 12, 2025

View reviewed changes

linjianchang closed this Oct 20, 2025

linjianchang force-pushed the master-36796 branch from 09e894a to 7c7a74d Compare October 20, 2025 09:12

linjianchang reopened this Oct 20, 2025

linjianchang closed this Oct 20, 2025

linjianchang reopened this Oct 20, 2025

lvyanquan self-assigned this Dec 15, 2025

lvyanquan added this to the V3.6.0 milestone Dec 25, 2025

lvyanquan reviewed Dec 25, 2025

View reviewed changes

lvyanquan reviewed Dec 26, 2025

View reviewed changes

linjianchang force-pushed the master-36796 branch 3 times, most recently from 08a8f61 to a6cabd0 Compare January 13, 2026 07:50

github-actions Bot removed the base label Jan 23, 2026

zml1206 reviewed Jan 29, 2026

View reviewed changes

linjc13 and others added 3 commits February 3, 2026 17:37

[FLINK-36796][pipeline-connector][oracle]add oracle pipeline connector.

c28d84d

Remove unnecessary config and update docs.

48d21d4

Try to remove unnecessary change.

64e4fd7

linjianchang force-pushed the master-36796 branch from 61e8a57 to 64e4fd7 Compare February 3, 2026 09:38

github-actions Bot added the build label Feb 3, 2026

lvyanquan added 2 commits February 6, 2026 13:01

Add oracle pipeline to test.

1c33a34

Use org.apache.flink.shaded.jackson2.com.fasterxml.jackson instead of…

e709237

… com.fasterxml.jackson.

lvyanquan approved these changes Feb 6, 2026

View reviewed changes

github-actions Bot added the approved label Feb 6, 2026

Update docs.

f6c8971

paulo-t reviewed Feb 6, 2026

View reviewed changes

Comment thread ...cle/src/main/java/org/apache/flink/cdc/connectors/oracle/source/OracleDataSourceOptions.java Outdated

Address comments.

3f7fa76

lvyanquan merged commit c8e444b into apache:master Feb 10, 2026
28 of 29 checks passed

paulo-t reviewed Feb 11, 2026

View reviewed changes

ThorneANN pushed a commit to ThorneANN/flink-cdc that referenced this pull request Mar 19, 2026

[FLINK-36796][pipeline-connector][oracle]add oracle pipeline connecto…

bde51ad

…r. (apache#3995)

Mrart pushed a commit to Mrart/flink-cdc that referenced this pull request Mar 26, 2026

[FLINK-36796][pipeline-connector][oracle]add oracle pipeline connecto…

ff07302

…r. (apache#3995)

sumitnekar reviewed Apr 4, 2026

View reviewed changes

		- `initial` （默认）：在第一次启动时对受监视的数据库表执行初始快照，并继续读取最新的 binlog。
		- `latest-offset`：首次启动时，从不对受监视的数据库表执行快照，连接器仅从 binlog 的结尾处开始读取，这意味着连接器只能读取在连接器启动之后的数据更改。

		- `initial` (default): Performs an initial snapshot on the monitored database tables upon first startup, and continue to read the latest binlog.
		- `latest-offset`: Never to perform snapshot on the monitored database tables upon first startup, just read from

		@@ -0,0 +1,26 @@
		################################################################################
		# Copyright 2023 Ververica Inc.

Conversation

linjianchang commented Apr 18, 2025

Uh oh!

joyCurry30 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joyCurry30 Apr 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

linjianchang commented Apr 23, 2025

Uh oh!

joyCurry30 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joyCurry30 Apr 19, 2025 •

edited

Loading