Amazon Redshift will no longer support the creation of new Python UDFs starting Patch 198. Existing Python UDFs will continue to function until June 30, 2026. For more information, see the blog post .

Use the Amazon Redshift driver metadata API for applications and tools

For applications and tools that connect to Amazon Redshift, such as a business intelligence tool or query editor, we recommend that you use the driver metadata API provided by the Amazon Redshift JDBC 2.x, ODBC 2.x, or Python drivers to discover metadata about your data warehouse objects, including databases, schemas, tables, columns, and data types. As alternatives, you can use Amazon Redshift SHOW commands.

Use the driver metadata API for the following benefits:

Example: Using JDBC DatabaseMetaData.getTables() to retrieve table metadata

DatabaseMetaData dbmd = connection.getMetaData();

// getTables(catalog, schemaPattern, tableNamePattern, types)
//   catalog:          "test" — filters to the database named "test"
//   schemaPattern:    "test_pattern" — filters schemas matching this pattern (supports SQL wildcards % and _)
//   tableNamePattern: null — no filter, returns all table names
//   types:            {"TABLE", "EXTERNAL TABLE"} — only return regular tables and external tables
ResultSet rs = dbmd.getTables("test", "test_pattern", null, new String[] {"TABLE", "EXTERNAL TABLE"});

Example: Using Python cursor.get_columns() to retrieve column metadata

cursor: redshift_connector.Cursor = conn.cursor()

# get_columns(catalog, schema_pattern, table_name_pattern, column_name_pattern)
#   catalog:             'test' — filters to the database named "test"
#   schema_pattern:      'test_pattern' — filters schemas matching this pattern (supports SQL wildcards % and _)
#   table_name_pattern:  'testabc' — filters to the table named "testabc"
#   column_name_pattern: '%' — wildcard, returns all columns in the matching table
result: tuple = cursor.get_columns('test', 'test_pattern', 'testabc', '%')

Example: Using ODBC SQLPrimaryKeys() to retrieve primary key metadata

// SQLPrimaryKeys(hstmt, catalog, catalog_len, schema, schema_len, table, table_len)
//   catalog: "test"         — filters to the database named "test"
//   schema:  "test_schema"  — filters to the schema named "test_schema"
//   table:   "test_table"   — retrieves primary key columns for this table
// Note: Unlike getTables/getColumns, SQLPrimaryKeys does NOT support wildcard patterns.
retcode = SQLPrimaryKeys(hstmt, (SQLCHAR *)"test", SQL_NTS, (SQLCHAR *)"test_schema", SQL_NTS, (SQLCHAR *)"test_table", SQL_NTS);

while (SQL_SUCCEEDED(retcode = SQLFetch(hstmt))) {
    for (i = 1; i <= columns; i++) {
        retcode = SQLGetData(hstmt, i, SQL_C_CHAR, buf, sizeof(buf), &indicator);
    }
}

Example: Using ODBC SQLTables() to list databases and schemas

The ODBC API does not provide separate functions for listing catalogs or schemas. Instead, you use special calling conventions of SQLTables() to retrieve this information.

To list all databases (catalogs)

Call SQLTables() with CatalogName set to SQL_ALL_CATALOGS. Set SchemaName and TableName to empty strings. The result set returns valid values only in the TABLE_CAT column. All other columns contain NULLs.

// List all catalogs (databases) available on the data source.
retcode = SQLTables(hstmt,
    (SQLCHAR *)SQL_ALL_CATALOGS, SQL_NTS, // CatalogName = "%" (SQL_ALL_CATALOGS)
    (SQLCHAR *)"", 0,                     // SchemaName = "" (empty string)
    (SQLCHAR *)"", 0,                     // TableName  = "" (empty string)
    NULL, 0);                             // TableType  = NULL (not filtered)

To list all schemas

Call SQLTables() with SchemaName set to SQL_ALL_SCHEMAS. Set CatalogName and TableName to empty strings.

// List all schemas available on the data source.
retcode = SQLTables(hstmt,
    (SQLCHAR *)"", 0,                     // CatalogName = "" (empty string)
    (SQLCHAR *)SQL_ALL_SCHEMAS, SQL_NTS,  // SchemaName = "%" (SQL_ALL_SCHEMAS)
    (SQLCHAR *)"", 0,                     // TableName  = "" (empty string)
    NULL, 0);                             // TableType  = NULL (not filtered)