apache/seatunnel · error · IllegalStateException

Cannot find table schema for table ${tableId}. Tried table i

Error message

Cannot find table schema for table ${tableId}. Tried table ids: ${tableIdWithoutCatalog} and ${tableIdWithCatalog}.

What it means

PostgresSnapshotReadTask.resolveTableSchema looks up the table's schema in the Debezium DatabaseSchema first by the user-supplied TableId (schema.table) and then with the connector catalog prepended (catalog.schema.table). If neither lookup returns a table, it throws this IllegalStateException because the snapshot reader cannot produce column metadata for the split.

Source

Thrown at seatunnel-connectors-v2/connector-cdc/connector-cdc-postgres/src/main/java/org/apache/seatunnel/connectors/seatunnel/cdc/postgres/source/reader/snapshot/PostgresSnapshotSplitReadTask.java:197

        String catalog = tableId.catalog();
        if (catalog == null || catalog.isEmpty()) {
            catalog = connectorConfig.databaseName();
        }
        if (catalog == null || catalog.isEmpty()) {
            throw new IllegalStateException(
                    String.format(
                            "Cannot find table schema for table %s. Tried table id: %s.",
                            tableId, tableIdWithoutCatalog));
        }

        TableId tableIdWithCatalog = new TableId(catalog, tableId.schema(), tableId.table());
        table = databaseSchema.tableFor(tableIdWithCatalog);
        if (table != null) {
            return table;
        }

        throw new IllegalStateException(
                String.format(
                        "Cannot find table schema for table %s. Tried table ids: %s and %s.",
                        tableId, tableIdWithoutCatalog, tableIdWithCatalog));
    }

    /** Dispatches the data change events for the records of a single table. */
    private void createDataEventsForTable(
            PostgresSnapshotContext snapshotContext,
            EventDispatcher.SnapshotReceiver snapshotReceiver,
            Table table)
            throws InterruptedException {

        long exportStart = clock.currentTimeInMillis();
        log.info("Exporting data from split '{}' of table {}", snapshotSplit.splitId(), table.id());

        final String selectSql =
                PostgresUtils.buildSplitScanQuery(
                        table,

View on GitHub (pinned to cf67b549a7)

Solutions

1. Verify the table exists and check exact spelling/case of database, schema and table in the config against \dt in psql
2. Confirm the table is matched by the table-names/database-name config and the connector's table filter so its schema is loaded into the DatabaseSchema at startup
3. Re-run the job (restart snapshot) after recreating/renaming the table so the schema cache is rebuilt
4. Check for concurrent DDL (DROP/RENAME/ALTER) during the snapshot and schedule snapshots outside DDL windows

Example fix

// before
table-names = ["public", "orders"]   // table actually named "Orders"
// after
table-names = ["public", "Orders"]   // match exact Postgres case

// before running the job, verify the table id resolves
String sql = "SELECT 1 FROM information_schema.tables WHERE table_schema=? AND table_name=?";
try (PreparedStatement ps = conn.prepareStatement(sql)) {
    ps.setString(1, schema); ps.setString(2, table);
    try (ResultSet rs = ps.executeQuery()) { if (!rs.next()) throw new IllegalStateException("table missing: " + schema + "." + table); }
}

try { snapshotReader.createDataEvents(split, records); } catch (IllegalStateException e) {
    if (e.getMessage().contains("Cannot find table schema")) { log.error("Schema missing for split table; re-enumerate and restart snapshot"); }
    throw e;
}

When it happens

Trigger: createDataEvents dispatches a snapshot split for a TableId that is absent from the DatabaseSchema built at task startup — e.g. the table was dropped or renamed between split enumeration and snapshot read, the tableId case does not match the actual Postgres schema/table name, or tableIdWithoutCatalog/tableIdWithCatalog normalization fails (catalog name differs from what the schema was built with).

Common situations: Case-sensitive table names quoted differently in the seatunnel config than in Postgres; dropping/truncating the table while a snapshot is running; running a job whose config lists a table not included in the CDC table filter so the schema was never loaded; a catalog configuration mismatch (database name changed).

Understand the failure class

Background: "Not found" and "does not exist" errors: why "Task not found", "No such folder", and "Can't find" fire when a lookup comes back empty — this error's family across 14 libraries.

Related errors

• Snapshotting of table ${table.id()} failed
• Can't obtain schema for table ${tableId}
• Snapshot was interrupted before completion
• Unsupported alter table event:
• Unsupported alter table event:

AI-assisted analysis of apache/seatunnel@cf67b549a7 (2026-09-10). Data as JSON: /api/errors/b27bdc49d685f687. Report an issue: GitHub.