CREATE EXTERNAL TABLE
Valid in: SQL, ESQL, OpenAPI, ODBC, JDBC, .NET
The CREATE EXTERNAL TABLE statement maps the structure of a data file created outside of Vector to the structure of a Vector table. The data can then be queried from its original locations.
This statement has the following format:
CREATE EXTERNAL TABLE [schema.]table_name
(column_name data_type {,column_name data_type})
USING provider_name
WITH REFERENCE='reference'
[,FORMAT='format']
[EXTERNAL_ROLE='external_role']
[,OPTIONS=('key'=value {,'key'='value'})]
table_name
Defines the name of the external table to be created.
column_name data_type
Specifies the column name and data type of each column. Each column specification must be separated with a comma.
USING provider_name
Specifies the name of the provider. The only valid provider is SPARK.
Provider Name Application
SPARK Spark
WITH REFERENCE='reference'
(Required) Specifies the reference to the external datasource. This corresponds to the parameter passed to the load method of DataFrameReader or save method of DataFrameWriter. For example:
For HDFS file: hdfs://namenode:port/path/to/file
For database systems: The table name
FORMAT='format'
(Optional) Is a WITH clause option that specifies the format of the external data.
Example formats include: csv, avro, parquet, hive, orc, json, jdbc. For other datasources, format corresponds to the class name that defines that external datasource. For example, for Redshift it would be com.databricks.spark.redshift. This corresponds to the parameter passed to the format method of DataFrameReader/Writer.
When FORMAT is not specified, the Spark-Vector Provider tries to recognize the format for HDFS files by looking at the file extension. For example, hdfs://namenode:port/tmp/my_file.orc is treated as an ORC file.
EXTERNAL_ROLE='external_role'
Note: This option is not implemented.
OPTIONS
(Optional) Is a WITH clause option that specifies user defined options for the datasource read or written to. This corresponds to the options method of the DataFrameReader/Writer. For example, for CSV files you can pass any options supported by spark-csv.