Data Types
In most cases, you don't need to create streams manually or specify the data type for the columns. Timeplus source will automatically create the streams/columns with proper types when you load data from Kafka/CSV/etc.
Like many analytics systems, the following common types are supported.
| Category | Type | Example | Note | Related functions |
|---|---|---|---|---|
| Numeric Types | integer | -100 | default with 4 bytes. You can also use int, smallint, bigint, or even uint16 etc. | to_int |
| decimal | 3.14 | decimal(precision, scale). Valid range for precision is [1: 76], valid range for scale is [0: precision] | to_decimal | |
| float | -3.1415 | default with 4 bytes. You can also use float64 or double for 8 bytes | to_float | |
| Boolean Type | bool | true | true or false | |
| String Type | string | "Hello" | strings of an arbitrary length. You can also use varchar To create string columns with fixed size in bytes, use fixed_string(positiveInt) | to_string, etc. |
| Universally Unique Identifier | uuid | 1f71acbf-59fc-427d-a634-1679b48029a9 | a universally unique identifier (UUID) is a 16-byte number used to identify records. For detailed information about the UUID, see Wikipedia | uuid |
| JSON Type | json | {"a":1} | a data type with built-in JSON support, with better query performance comparing saving the JSON as string and extracting value at query time. Suitable for JSON documents in same schema. Access the value via column.a (instead of column:a for query-time JSON extraction) | |
| Date and Time Types | date | '2022-05-16' | without time | to_date, today |
| datetime | '2022-05-16 11:01:02' | with second | to_time, now | |
| datetime64 | '2022-05-16 11:01:02.345' | with millisecond, same as datetime64(3) | to_time, now64 | |
| Compound Types | array | [1,2] | access 1st element via array[1] | length, array_concat |
| map | map_cast('k1','v1','k2','v2') | access key1 via map['key1'] | map_cast | |
| tuple | (1,2) | access 1st element via tuple.1 | tuple_cast |