Primitive Data Types

Table 1 lists the primitive data types supported by DLI.

**Table 1** Primitive data types
Data Type	Description	Storage Space	Value Range	Support by OBS Table	Support by DLI Table
INT	Signed integer	4 bytes	–2147483648 to 2147483647	Yes	Yes
STRING	String	-	-	Yes	Yes
FLOAT	Single-precision floating point	4 bytes	-	Yes	Yes
DOUBLE	Double-precision floating-point	8 bytes	-	Yes	Yes
DECIMAL(precision,scale)	Decimal number. Data type of valid fixed places and decimal places, for example, 3.5. precision: indicates the maximum number of digits that can be displayed. scale: indicates the number of decimal places.	-	1<=precision<=38 0<=scale<=38 If precision and scale are not specified, DECIMAL (38,38) is used by default.	Yes	Yes
BOOLEAN	Boolean	1 byte	TRUE/FALSE	Yes	Yes
SMALLINT/SHORT	Signed integer	2 bytes	-32768~32767	Yes	Yes
TINYINT	Signed integer	1 byte	-128~127	Yes	No
BIGINT/LONG	Signed integer	8 bytes	–9223372036854775808 to 9223372036854775807	Yes	Yes
TIMESTAMP	Timestamp in raw data format, indicating the date and time Example: 1621434131222	-	-	Yes	Yes
CHAR	Fixed-length string	-	-	Yes	Yes
VARCHAR	Variable-length string	-	-	Yes	Yes
DATE	Date type in the format of yyyy-mm-dd, for example, 2014-05-29	-	DATE does not contain time information. Its value ranges from 0000-01-01 to 9999-12-31.	Yes	Yes

VARCHAR and CHAR data is stored in STRING type on DLI. Therefore, the string that exceeds the specified length will not be truncated.
FLOAT data is stored as DOUBLE data on DLI.

INT

Signed integer with a storage space of 4 bytes. Its value ranges from –2147483648 to 2147483647. If this field is NULL, value 0 is used by default.

STRING

String.

FLOAT

Single-precision floating point with a storage space of 4 bytes. If this field is NULL, value 0 is used by default.

Due to the limitation of storage methods of floating point data, do not use the formula a==b to check whether two floating point values are the same. You are advised to use the formula: absolute value of (a-b) <= EPSILON. EPSILON indicates the allowed error range which is usually 1.19209290E-07F. If the formula is satisfied, the compared two floating point values are considered the same.

DOUBLE

Double-precision floating point with a storage space of 8 bytes. If this field is NULL, value 0 is used by default.

Due to the limitation of storage methods of floating point data, do not use the formula a==b to check whether two floating point values are the same. You are advised to use the formula: absolute value of (a-b) <= EPSILON. EPSILON indicates the allowed error range which is usually 2.2204460492503131E-16. If the formula is satisfied, the compared two floating point values are considered the same.

DECIMAL

Decimal(p,s) indicates that the total digit length is p, including p – s integer digits and s fractional digits. p indicates the maximum number of decimal digits that can be stored, including the digits to both the left and right of the decimal point. The value of p ranges from 1 to 38. s indicates the maximum number of decimal digits that can be stored to the right of the decimal point. The fractional digits must be values ranging from 0 to p. The fractional digits can be specified only after significant digits are specified. Therefore, the following inequality is concluded: 0 ≤ s ≤ p. For example, decimal (10,6) indicates that the value contains 10 digits, in which there are four integer digits and six fractional digits.

BOOLEAN

Boolean, which can be TRUE or FALSE.

SMALLINT/SHORT

Signed integer with a storage space of 2 bytes. Its value ranges from –32768 to 32767. If this field is NULL, value 0 is used by default.

TINYINT

Signed integer with a storage space of 1 byte. Its value ranges from –128 to 127. If this field is NULL, value 0 is used by default.

BIGINT/LONG

Signed integer with a storage space of 8 bytes. Its value ranges from –9223372036854775808 to 9223372036854775807. It does not support scientific notation. If this field is NULL, value 0 is used by default.

TIMESTAMP

Legacy UNIX TIMESTAMP is supported, providing the precision up to the microsecond level. TIMESTAMP is defined by the difference between the specified time and UNIX epoch (UNIX epoch time: 1970-01-01 00:00:00) in seconds. The data type STRING can be implicitly converted to TIMESTAMP, but it must be in the yyyy-MM-dd HH:mm:SS[.ffffff] format. The precision after the decimal point is optional.)

CHAR

String with a fixed length. In DLI, the STRING type is used.

VARCHAR

VARCHAR is declared with a length that indicates the maximum number of characters in a string. During conversion from STRING to VARCHAR, if the number of characters in STRING exceeds the specified length, the excess characters of STRING are automatically trimmed. Similar to STRING, the spaces at the end of VARCHAR are meaningful and affect the comparison result. In DLI, the STRING type is used.

DATE

DATE supports only explicit conversion (cast) with DATE, TIMESTAMP, and STRING. For details, see Table 2.

**Table 2** cast function conversion
Explicit Conversion	Conversion Result
cast(date as date)	Same as value of DATE.
cast(timestamp as date)	The date (yyyy-mm-dd) is obtained from TIMESTAMP based on the local time zone and returned as the value of DATE.
cast(string as date)	If the STRING is in the yyyy-MM-dd format, the corresponding date (yyyy-mm-dd) is returned as the value of DATE. If the STRING is not in the yyyy-MM-dd format, NULL is returned.
cast(date as timestamp)	Timestamp that maps to the zero hour of the date (yyyy-mm-dd) specified by DATE is generated based on the local time zone and returned as the value of DATE.
cast(date as string)	A STRING in the yyyy-MM-dd format is generated based on the date (yyyy-mm-dd) specified by DATE and returned as the value of DATE.