CREATE TABLE

Creates a new table.

Syntax

CREATE TABLE [schema_name.]table_name 
(
column_name <data_type> [PRIMARY KEY] [,...]
)
[PARTITION BY RANGE(column_name) EVERY (<part_interval>)]

<Data_type> ::=
{
INT | BIGINT | FLOAT | DOUBLE | STRING | TIMESTAMP
}

<part_interval> ::=
{
n      | 
VALUE  | 
INTERVAL 'n' {month | day | hour}
}


CREATE TABLE [schema_name.]table_name MERGE FROM [schema_name.]source-table1, [schema_name.]source-table2[,...]

Description

Create a new table object. Table name and column names must be a valid identifier. For more details on Jethro data types attributes, see Data Types. A table can have one primary key column. Primary keys must contain unique values and cannot contain NULL values.

Parameter Details

Parameter

Mandatory?

Details

PARTITION BY

OPTIONAL

The clause defines range partitioning.

Unlike Hive and Impala, in Jethro the partition column name must be one of the columns previously defined in the table column list.

Partitioning interval specification depends on the data type of the partitioning column:

For numeric partition key - use EVERY (#n).
- For example: PARTITION BY RANGE(store_sk) EVERY (10)

For string partition key - use EVERY (VALUE) without replacing the 'value' with anything. The result would be that each distinct string will have its own partition.
- For example: PARTITION BY RANGE(store_name) EVERY (VALUE)

For timestamp partition key - use EVERY (INTERVAL 'n'...)
- For example: PARTITION BY RANGE(sale_date) EVERY (INTERVAL '3' MONTH)

*The valid range for 'n' is: 0<n<(2^32-1), integer values only.

Merged Table

Table can also be created via vertical merge of two or more tables using the MERGE FROM clause. Merged table will hold all the columns from all it's source tables.

Description

Parameter Details

Merged Table

Merge Table Constraints:

See Also