[FLINK-38779] Move TableIdRouter to flink-cdc-common package

Author: yuxiqian

…erators/schema/common/TableIdRouter.java …link/cdc/common/route/TableIdRouter.javaflink-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/common/TableIdRouter.java renamed to flink-cdc-common/src/main/java/org/apache/flink/cdc/common/route/TableIdRouter.java flink-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/common/TableIdRouter.java renamed to flink-cdc-common/src/main/java/org/apache/flink/cdc/common/route/TableIdRouter.java

@@ -15,11 +15,12 @@
  * limitations under the License.
  */
 
-package org.apache.flink.cdc.runtime.operators.schema.common;
+package org.apache.flink.cdc.common.route;
 
 import org.apache.flink.api.java.tuple.Tuple3;
+import org.apache.flink.cdc.common.annotation.PublicEvolving;
 import org.apache.flink.cdc.common.event.TableId;
-import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.common.schema.Selectors;
 
 import org.apache.flink.shaded.guava31.com.google.common.cache.CacheBuilder;
 import org.apache.flink.shaded.guava31.com.google.common.cache.CacheLoader;
@@ -44,6 +45,7 @@
  * Calculates how upstream data change events should be dispatched to downstream tables. Returns one
  * or many destination Table IDs based on provided routing rules.
  */
+@PublicEvolving
 public class TableIdRouter {
 
     private static final Logger LOG = LoggerFactory.getLogger(TableIdRouter.class);
@@ -54,7 +56,21 @@ public class TableIdRouter {
 
     private static final String DOT_PLACEHOLDER = "_dot_placeholder_";
 
-    private static Pattern validateTableListToRegExpPattern(String tables) {
+    /**
+     * Currently, The supported regular syntax is not exactly the same in {@link Selectors}.
+     *
+     * <p>The main discrepancies are :
+     *
+     * <p>1) {@link Selectors} use {@code ,} to split table names instead of `|`.
+     *
+     * <p>2) If there is a need to use a dot ({@code .}) in a regular expression to match any
+     * character, it is necessary to escape the dot with a backslash.
+     *
+     * <p>3) The unescaped {@code .} is used as the separator of database and table name. When
+     * converting to Debezium style, it is expected to be escaped to match the dot ({@code .})
+     * literally instead of the meta-character.
+     */
+    public static String convertTableListToRegExpPattern(String tables) {
         LOG.info("Rewriting CDC style table capture list: {}", tables);
 
         // In CDC-style table matching, table names could be separated by `,` character.
@@ -86,7 +102,7 @@ private static Pattern validateTableListToRegExpPattern(String tables) {
                 unescapedTablesWithDbTblSeparator.replace(DOT_PLACEHOLDER, ".");
         LOG.info("Final standard RegExp table capture list: {}", standardRegExpTableCaptureList);
 
-        return Pattern.compile(standardRegExpTableCaptureList);
+        return standardRegExpTableCaptureList;
     }
 
     public TableIdRouter(List<RouteRule> routingRules) {
@@ -95,7 +111,7 @@ public TableIdRouter(List<RouteRule> routingRules) {
             try {
                 routes.add(
                         new Tuple3<>(
-                                validateTableListToRegExpPattern(rule.sourceTable),
+                                Pattern.compile(convertTableListToRegExpPattern(rule.sourceTable)),
                                 rule.sinkTable,
                                 rule.replaceSymbol));
             } catch (PatternSyntaxException e) {

flink-cdc-connect/flink-cdc-pipeline-connectors/flink-cdc-pipeline-connector-mysql/src/main/java/org/apache/flink/cdc/connectors/mysql/factory/MySqlDataSourceFactory.java

@@ -30,7 +30,6 @@
 import org.apache.flink.cdc.common.source.DataSource;
 import org.apache.flink.cdc.common.utils.StringUtils;
 import org.apache.flink.cdc.connectors.mysql.source.MySqlDataSource;
-import org.apache.flink.cdc.connectors.mysql.source.MySqlDataSourceOptions;
 import org.apache.flink.cdc.connectors.mysql.source.config.MySqlSourceConfigFactory;
 import org.apache.flink.cdc.connectors.mysql.source.config.ServerIdRange;
 import org.apache.flink.cdc.connectors.mysql.source.offset.BinlogOffset;
@@ -45,7 +44,6 @@
 
 import com.mysql.cj.conf.PropertyKey;
 import io.debezium.relational.RelationalDatabaseConnectorConfig;
-import io.debezium.relational.Tables;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
@@ -63,6 +61,7 @@
 import java.util.Set;
 import java.util.stream.Collectors;
 
+import static org.apache.flink.cdc.common.route.TableIdRouter.convertTableListToRegExpPattern;
 import static org.apache.flink.cdc.connectors.mysql.source.MySqlDataSourceOptions.CHUNK_KEY_EVEN_DISTRIBUTION_FACTOR_LOWER_BOUND;
 import static org.apache.flink.cdc.connectors.mysql.source.MySqlDataSourceOptions.CHUNK_KEY_EVEN_DISTRIBUTION_FACTOR_UPPER_BOUND;
 import static org.apache.flink.cdc.connectors.mysql.source.MySqlDataSourceOptions.CHUNK_META_GROUP_SIZE;
@@ -231,7 +230,7 @@ public DataSource createDataSource(Context context) {
         }
 
         if (scanBinlogNewlyAddedTableEnabled) {
-            String newTables = validateTableAndReturnDebeziumStyle(tables);
+            String newTables = convertTableListToRegExpPattern(tables);
             configFactory.tableList(newTables);
             configFactory.excludeTableList(tablesExclude);
 
@@ -516,58 +515,6 @@ && doubleCompare(distributionFactorLower, 1.0d) <= 0,
 
     private static final String DOT_PLACEHOLDER = "_$dot_placeholder$_";
 
-    /**
-     * Currently, The supported regular syntax is not exactly the same in {@link Selectors} and
-     * {@link Tables.TableFilter}.
-     *
-     * <p>The main distinction are :
-     *
-     * <p>1) {@link Selectors} use {@code ,} to split table names and {@link Tables.TableFilter} use
-     * `|` to split table names.
-     *
-     * <p>2) If there is a need to use a dot ({@code .}) in a regular expression to match any
-     * character, it is necessary to escape the dot with a backslash, refer to {@link
-     * MySqlDataSourceOptions#TABLES}.
-     *
-     * <p>3) The unescaped {@code .} is used as the separator of database and table name. When
-     * converting to Debezium style, it is expected to be escaped to match the dot ({@code .})
-     * literally instead of the meta-character.
-     */
-    private String validateTableAndReturnDebeziumStyle(String tables) {
-        LOG.info("Rewriting CDC style table capture list: {}", tables);
-
-        // In CDC-style table matching, table names could be separated by `,` character.
-        // Convert it to `|` as it's standard RegEx syntax.
-        tables =
-                Arrays.stream(tables.split(",")).map(String::trim).collect(Collectors.joining("|"));
-        LOG.info("Expression after replacing comma with vert separator: {}", tables);
-
-        // Essentially, we're just trying to swap escaped `\\.` and unescaped `.`.
-        // In our table matching syntax, `\\.` means RegEx token matcher and `.` means database &
-        // table name separator.
-        // On the contrary, while we're matching TableId string, `\\.` means matching the "dot"
-        // literal and `.` is the meta-character.
-
-        // Step 1: escape the dot with a backslash, but keep it as a placeholder (like `$`).
-        // For example, `db\.*.tbl\.*` => `db$*.tbl$*`
-        String unescapedTables = tables.replace("\\.", DOT_PLACEHOLDER);
-        LOG.info("Expression after unescaping dots as RegEx meta-character: {}", unescapedTables);
-
-        // Step 2: replace all remaining dots (`.`) to quoted version (`\.`), as a separator between
-        // database and table names.
-        // For example, `db$*.tbl$*` => `db$*\.tbl$*`
-        String unescapedTablesWithDbTblSeparator = unescapedTables.replace(".", "\\.");
-        LOG.info("Re-escaping dots as TableId delimiter: {}", unescapedTablesWithDbTblSeparator);
-
-        // Step 3: restore placeholder to normal RegEx matcher (`.`)
-        // For example, `db$*\.tbl$*` => `db.*\.tbl.*`
-        String debeziumStyleTableCaptureList =
-                unescapedTablesWithDbTblSeparator.replace(DOT_PLACEHOLDER, ".");
-        LOG.info("Final Debezium-style table capture list: {}", debeziumStyleTableCaptureList);
-
-        return debeziumStyleTableCaptureList;
-    }
-
     /** Replaces the default timezone placeholder with session timezone, if applicable. */
     private static ZoneId getServerTimeZone(Configuration config) {
         final String serverTimeZone = config.get(SERVER_TIME_ZONE);

flink-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/common/SchemaDerivator.java

@@ -28,6 +28,7 @@
 import org.apache.flink.cdc.common.event.SchemaChangeEventWithPreSchema;
 import org.apache.flink.cdc.common.event.TableId;
 import org.apache.flink.cdc.common.pipeline.SchemaChangeBehavior;
+import org.apache.flink.cdc.common.route.TableIdRouter;
 import org.apache.flink.cdc.common.schema.Column;
 import org.apache.flink.cdc.common.schema.Schema;
 import org.apache.flink.cdc.common.sink.MetadataApplier;

flink-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/common/SchemaRegistry.java

@@ -22,6 +22,7 @@
 import org.apache.flink.cdc.common.event.TableId;
 import org.apache.flink.cdc.common.pipeline.SchemaChangeBehavior;
 import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.common.route.TableIdRouter;
 import org.apache.flink.cdc.common.schema.Schema;
 import org.apache.flink.cdc.common.sink.MetadataApplier;
 import org.apache.flink.cdc.runtime.operators.schema.common.event.FlushSuccessEvent;

flink-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/distributed/SchemaOperator.java

@@ -26,11 +26,11 @@
 import org.apache.flink.cdc.common.exceptions.SchemaEvolveException;
 import org.apache.flink.cdc.common.pipeline.SchemaChangeBehavior;
 import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.common.route.TableIdRouter;
 import org.apache.flink.cdc.common.schema.Schema;
 import org.apache.flink.cdc.common.utils.SchemaUtils;
 import org.apache.flink.cdc.runtime.operators.schema.common.CoordinationResponseUtils;
 import org.apache.flink.cdc.runtime.operators.schema.common.SchemaDerivator;
-import org.apache.flink.cdc.runtime.operators.schema.common.TableIdRouter;
 import org.apache.flink.cdc.runtime.operators.schema.common.metrics.SchemaOperatorMetrics;
 import org.apache.flink.cdc.runtime.operators.schema.distributed.event.SchemaChangeRequest;
 import org.apache.flink.cdc.runtime.operators.schema.distributed.event.SchemaChangeResponse;

flink-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/regular/BatchSchemaOperator.java

@@ -25,11 +25,11 @@
 import org.apache.flink.cdc.common.event.TableId;
 import org.apache.flink.cdc.common.pipeline.SchemaChangeBehavior;
 import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.common.route.TableIdRouter;
 import org.apache.flink.cdc.common.schema.Schema;
 import org.apache.flink.cdc.common.sink.MetadataApplier;
 import org.apache.flink.cdc.runtime.operators.schema.common.SchemaDerivator;
 import org.apache.flink.cdc.runtime.operators.schema.common.SchemaManager;
-import org.apache.flink.cdc.runtime.operators.schema.common.TableIdRouter;
 import org.apache.flink.streaming.api.graph.StreamConfig;
 import org.apache.flink.streaming.api.operators.AbstractStreamOperator;
 import org.apache.flink.streaming.api.operators.ChainingStrategy;

flink-cdc-runtime/src/main/java/org/apache/flink/cdc/runtime/operators/schema/regular/SchemaOperator.java

@@ -26,11 +26,11 @@
 import org.apache.flink.cdc.common.event.TableId;
 import org.apache.flink.cdc.common.pipeline.SchemaChangeBehavior;
 import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.common.route.TableIdRouter;
 import org.apache.flink.cdc.common.schema.Schema;
 import org.apache.flink.cdc.common.utils.SchemaUtils;
 import org.apache.flink.cdc.runtime.operators.schema.common.CoordinationResponseUtils;
 import org.apache.flink.cdc.runtime.operators.schema.common.SchemaDerivator;
-import org.apache.flink.cdc.runtime.operators.schema.common.TableIdRouter;
 import org.apache.flink.cdc.runtime.operators.schema.common.metrics.SchemaOperatorMetrics;
 import org.apache.flink.cdc.runtime.operators.schema.regular.event.SchemaChangeRequest;
 import org.apache.flink.cdc.runtime.operators.schema.regular.event.SchemaChangeResponse;

…ors/schema/common/TableIdRouterTest.java …/cdc/common/route/TableIdRouterTest.javaflink-cdc-runtime/src/test/java/org/apache/flink/cdc/runtime/operators/schema/common/TableIdRouterTest.java renamed to flink-cdc-runtime/src/test/java/org/apache/flink/cdc/common/route/TableIdRouterTest.java flink-cdc-runtime/src/test/java/org/apache/flink/cdc/runtime/operators/schema/common/TableIdRouterTest.java renamed to flink-cdc-runtime/src/test/java/org/apache/flink/cdc/common/route/TableIdRouterTest.java

@@ -15,10 +15,10 @@
  * limitations under the License.
  */
 
-package org.apache.flink.cdc.runtime.operators.schema.common;
+package org.apache.flink.cdc.common.route;
 
 import org.apache.flink.cdc.common.event.TableId;
-import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.runtime.operators.schema.common.SchemaTestBase;
 
 import org.junit.jupiter.api.Test;
 
@@ -40,6 +40,17 @@ private static List<String> testRoute(String tableId) {
                 .collect(Collectors.toList());
     }
 
+    private static String testConvert(String input) {
+        return TableIdRouter.convertTableListToRegExpPattern(input);
+    }
+
+    @Test
+    void testConvertingDebeziumTableIdToStandardRegex() {
+        assertThat(testConvert("foo.bar")).isEqualTo("foo\\.bar");
+        assertThat(testConvert("foo.bar,foo.baz")).isEqualTo("foo\\.bar|foo\\.baz");
+        assertThat(testConvert("db.\\.*")).isEqualTo("db\\..*");
+    }
+
     @Test
     void testImplicitRoute() {
         assertThat(testRoute("db_0.table_1")).containsExactlyInAnyOrder("db_0.table_1");

flink-cdc-runtime/src/test/java/org/apache/flink/cdc/runtime/operators/schema/common/SchemaDerivatorTest.java

@@ -28,6 +28,7 @@
 import org.apache.flink.cdc.common.event.TruncateTableEvent;
 import org.apache.flink.cdc.common.pipeline.SchemaChangeBehavior;
 import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.common.route.TableIdRouter;
 import org.apache.flink.cdc.common.schema.Column;
 import org.apache.flink.cdc.common.schema.Schema;
 import org.apache.flink.cdc.common.sink.MetadataApplier;

flink-cdc-runtime/src/test/java/org/apache/flink/cdc/runtime/operators/schema/common/SchemaTestBase.java

@@ -23,6 +23,7 @@
 import org.apache.flink.cdc.common.event.Event;
 import org.apache.flink.cdc.common.event.TableId;
 import org.apache.flink.cdc.common.route.RouteRule;
+import org.apache.flink.cdc.common.route.TableIdRouter;
 import org.apache.flink.cdc.common.types.DataType;
 import org.apache.flink.cdc.common.types.DataTypes;
 import org.apache.flink.cdc.runtime.partitioning.PartitioningEvent;