Define the name of the database which will store the data.

mysql> SET @db_list = '["tpch"]';

This assumes that Lakehouse Auto Parallel Load will analyze the data, infer the table structure, and create the database and all tables.

Define the db_object parameters that will load data from three external sources with Avro, CSV and Parquet format files:

mysql> SET @ext_tables = '[{
  "db_name": "tpch",
  "tables": [{
    "table_name": "supplier_pq",
    "dialect": {"format": "parquet"},
    "file": [{
      "prefix": "src_data/parquet/tpch/supplier/",
      "bucket": "myBucket",
      "namespace": "myNamespace",
      "region": "myRegion"}]
  },
  {
    "table_name": "nation_csv",
    "dialect": {
      "format": "csv",
      "field_delimiter": "|",
      "record_delimiter": "|\\n",
      "has_header": true
    },
    "file": [{
      "par": "https://objectstorage.../nation.csv"}]
  },
  {
    "table_name": "region_avro",
    "dialect": {"format": "avro"},
    "file": [{
      "par": "https://objectstorage.../region.avro"}]
  }]
}]';

Define the @options variable with SET:

mysql> SET @options = JSON_OBJECT('external_tables', CAST(@ext_tables AS JSON));

Setting mode to dryrun generates the load script but does not create or load the external tables. For example:

mysql> SET @options = JSON_OBJECT('mode', 'dryrun', 'external_tables', CAST(@ext_tables AS JSON));

To implement the changes as part of the load command, set mode to normal. This is the default, and it is not necessary to add it to the command.