Numeric Types

Table 1 lists all available types. For arithmetic operators and related built-in functions, see Arithmetic Functions and Operators.

**Table 1** Integer types
Name	Description	Storage Space	Range
TINYINT	Tiny integer, also called INT1	1 byte	0 ~ +255
SMALLINT	Small integer, also called INT2	2 bytes	–32,768 to +32,767
INTEGER	Typical choice for integers, also called INT4	4 bytes	–2,147,483,648 to +2,147,483,647
BINARY_INTEGER	Alias of INTEGER.	4 bytes	–2,147,483,648 to +2,147,483,647
BIGINT	Big integer, also called INT8	8 bytes	–9,223,372,036,854,775,808 to +9,223,372,036,854,775,807
int16	A 16-byte integer cannot be used to create tables.	16 bytes	–170,141,183,460,469,231,731,687,303,715,884,105,728 to +170,141,183,460,469,231,731,687,303,715,884,105,727

Example:

   
        -- Create a table containing TINYINT data.
gaussdb=# CREATE TABLE int_type_t1
           (
            IT_COL1 TINYINT
           );

-- Insert data to the created table.
gaussdb=# INSERT INTO int_type_t1 VALUES(10);

-- View data.
gaussdb=# SELECT * FROM int_type_t1;
 it_col1  
--------- 
 10
(1 row)

-- Drop the table.
gaussdb=# DROP TABLE int_type_t1;

   
        -- Create a table containing TINYINT, INTEGER, and BIGINT data.
gaussdb=# CREATE TABLE int_type_t2 
(
    a TINYINT, 
    b TINYINT,
    c INTEGER,
    d BIGINT
);

-- Insert data.
gaussdb=# INSERT INTO int_type_t2 VALUES(100, 10, 1000, 10000);

-- View data.
gaussdb=# SELECT * FROM int_type_t2;
  a  | b  |  c   |   d   
-----+----+------+-------
 100 | 10 | 1000 | 10000
(1 row)

-- Drop the table.
gaussdb=# DROP TABLE int_type_t2;

Only numbers of the TINYINT, SMALLINT, INTEGER, BIGINT, or INT16 type, that is, integers can be stored. Saving a number with a decimal in any of the data types will result in errors.
The INTEGER type is the common choice, as it offers the best balance between range, storage size, and performance. Generally, use the SMALLINT type only if you are sure that the value range is within the SMALLINT value range. The storage speed of INTEGER is much faster. BIGINT is used only when the range of INTEGER is not large enough.

**Table 2** Arbitrary precision types
Name	Description	Storage Space	Range
NUMERIC[(p[,s])], DECIMAL[(p[,s])]	The value range of p is [1,1000], and the value range of s is [0,p]. NOTE: p indicates the total digits, and s indicates the decimal digit.	The precision is specified by users. Every four decimal digits occupy two bytes, and an extra eight-byte overhead is added to the entire data.	Up to 131,072 digits before the decimal point, and up to 16,383 digits after the decimal point when no precision is specified.
NUMBER[(p[,s])]	Alias of the NUMERIC type.	The precision is specified by users. Every four decimal digits occupy two bytes, and an extra eight-byte overhead is added to the entire data.	Up to 131,072 digits before the decimal point, and up to 16,383 digits after the decimal point when no precision is specified.

Example:

-- Create a table.
gaussdb=# CREATE TABLE decimal_type_t1 
(
    DT_COL1 DECIMAL(10,4)
);

-- Insert data.
gaussdb=# INSERT INTO decimal_type_t1 VALUES(123456.122331);

-- Query data in the table.
gaussdb=# SELECT * FROM decimal_type_t1;
   dt_col1   
-------------
 123456.1223
(1 row)

-- Drop the table.
gaussdb=# DROP TABLE decimal_type_t1;

   
        -- Create a table.
gaussdb=# CREATE TABLE numeric_type_t1 
(
    NT_COL1 NUMERIC(10,4)
);

-- Insert data.
gaussdb=# INSERT INTO numeric_type_t1 VALUES(123456.12354);

-- Query data in the table.
gaussdb=# SELECT * FROM numeric_type_t1;
   nt_col1   
-------------
 123456.1235
(1 row)

-- Drop the table.
gaussdb=# DROP TABLE numeric_type_t1;

Compared to the integer types, the arbitrary precision numbers require larger storage space and have lower storage efficiency, operation efficiency, and poorer compression ratio results. The INTEGER type is the common choice when number types are defined. Arbitrary precision numbers are used when numbers exceed the maximum range indicated by the integers.
When NUMERIC/DECIMAL is used for defining a column, you are advised to specify the precision (p) and scale (s) for the column.

**Table 3** Sequence integer
Name	Description	Storage Space	Range
SMALLSERIAL	Two-byte serial integer	2 bytes.	–32,768 to +32,767.
SERIAL	Four-byte serial integer	4 bytes.	–2,147,483,648 to +2,147,483,647.
BIGSERIAL	Eight-byte serial integer	8 bytes.	–9,223,372,036,854,775,808 to +9,223,372,036,854,775,807.
LARGESERIAL	By default, a 16-byte auto-incrementing integer is inserted. The actual value type is the same as that of NUMERIC.	Variable-length type. Every four decimal digits occupy two bytes, and an extra eight-byte overhead is added to the entire data.	There can be a maximum of 131072 digits before the decimal point and 16383 digits after the decimal point.

Example:

   
    
      
      
        -- Create a table.
gaussdb=# CREATE TABLE smallserial_type_tab(a SMALLSERIAL);

-- Insert data.
gaussdb=# INSERT INTO smallserial_type_tab VALUES(default);

-- Insert data again.
gaussdb=# INSERT INTO smallserial_type_tab VALUES(default);

-- View data.
gaussdb=# SELECT * FROM smallserial_type_tab;  
 a 
---
 1
 2
(2 rows)

-- Create a table.
gaussdb=# CREATE TABLE serial_type_tab(b SERIAL);

-- Insert data.
gaussdb=# INSERT INTO serial_type_tab VALUES(default);

-- Insert data again.
gaussdb=# INSERT INTO serial_type_tab VALUES(default);

-- View data.
gaussdb=# SELECT * FROM serial_type_tab; 
 b 
---
 1
 2
(2 rows)

-- Create a table.
gaussdb=# CREATE TABLE bigserial_type_tab(c BIGSERIAL);

-- Insert data.
gaussdb=# INSERT INTO bigserial_type_tab VALUES(default);

-- Insert data again.
gaussdb=# INSERT INTO bigserial_type_tab VALUES(default);

-- View data.
gaussdb=# SELECT * FROM bigserial_type_tab;
 c 
---
 1
 2
(2 rows)

-- Create a table.
gaussdb=# CREATE TABLE largeserial_type_tab(c LARGESERIAL);

-- Insert data.
gaussdb=# INSERT INTO largeserial_type_tab VALUES(default);

-- Insert data again.
gaussdb=# INSERT INTO largeserial_type_tab VALUES(default);

-- View data.
gaussdb=# SELECT * FROM largeserial_type_tab;
 c 
---
 1
 2
(2 rows)

-- Drop the table.
gaussdb=# DROP TABLE smallserial_type_tab;

gaussdb=# DROP TABLE serial_type_tab;

gaussdb=# DROP TABLE bigserial_type_tab;

       

     

   
  

SMALLSERIAL, SERIAL, BIGSERIAL, and LARGESERIAL are not real types. They are concepts used for setting a unique identifier for a table. Therefore, an integer column is created and its default value plans to be read from a sequencer. A NOT NULL constraint is used to ensure NULL is not inserted. In most cases you would also want to attach a UNIQUE or PRIMARY KEY constraint to prevent duplicate values from being inserted unexpectedly, but this is not automatic. Finally, the sequencer belongs to the column. In this case, when the column or the table is deleted, the sequencer is also deleted. Currently, you can specify a SERIAL column when creating a table or add a SERIAL column to an ordinary table in PG-compatible mode. In addition, SERIAL columns cannot be created in temporary tables. Because SERIAL is not a data type, columns cannot be converted to this type.

**Table 4** Floating point types
Name	Description	Storage Space	Range
REAL, FLOAT4	Single precision floating points, which is not very precise.	4 bytes.	–3.402E+38 to +3.402E+38, 6-digit decimal digits.
DOUBLE PRECISION, FLOAT8	Double precision floating points, which is not very precise.	8 bytes.	–1.79E+308 to +1.79E+308, 15-bit decimal digits.
FLOAT[(p)]	Floating-point number, which is not very precise. The value range of precision (p) is [1,53].	4 bytes or 8 bytes.	REAL or DOUBLE PRECISION is selected as an internal identifier based on precision (p). If no precision is specified, DOUBLE PRECISION is used as the internal identifier.
BINARY_DOUBLE	Alias for DOUBLE PRECISION, compatible with Oracle.	8 bytes.	–1.79E+308 to +1.79E+308, 15-bit decimal digits.
DEC[(p[,s])]	The value range of p is [1,1000], and the value range of s is [0,p].	The precision is specified by users. Every four decimal digits occupy two bytes, and an extra eight-byte overhead is added to the entire data.	Maximum 131,072 digits before the decimal point and 16,383 digits after the decimal point when the precision and scale are specified to the maximum.
INTEGER[(p[,s])]	The value range of p (precision) is [1,1000], and the value range of s (scale) is [0,p]. If the precision and scale are not specified, the precision p is 10 and the scale s is 0 by default. If the precision and scale are not specified, this type is mapped to INTEGER. If the precision and scale are specified, this type is mapped to NUMERIC.	The precision is specified by users. Every four decimal digits occupy two bytes, and an extra eight-byte overhead is added to the entire data.	Maximum 131,072 digits before the decimal point and 16,383 digits after the decimal point when the precision and scale are specified to the maximum. If the precision and scale are not specified, the value ranges from –2,147,483,648 to +2,147,483,647.

In Table 4, p is the precision, indicating the minimum acceptable total number of integral places, and s indicates the decimal digit.

Example:

   
    
      
      
        -- Create a table.
gaussdb=# CREATE TABLE float_type_t2 
(
    FT_COL1 INTEGER,
    FT_COL2 FLOAT4,
    FT_COL3 FLOAT8,
    FT_COL4 FLOAT(3),
    FT_COL5 BINARY_DOUBLE,
    FT_COL6 DECIMAL(10,4),
    FT_COL7 INTEGER(6,3)
);

-- Insert data.
gaussdb=# INSERT INTO float_type_t2 VALUES(10,10.365456,123456.1234,10.3214, 321.321, 123.123654, 123.123654);

-- View data.
gaussdb=# SELECT * FROM float_type_t2 ;
 ft_col1 | ft_col2 |   ft_col3   | ft_col4 | ft_col5 | ft_col6  | ft_col7 
---------+---------+-------------+---------+---------+----------+---------
      10 | 10.3655 | 123456.1234 | 10.3214 | 321.321 | 123.1237 | 123.124
(1 row)

-- Drop the table.
gaussdb=# DROP TABLE float_type_t2;

       

     

   
  