Operators are represented by special characters or keywords, they do not use function call syntax.
Operator Precedence
The following table lists all SQL operators from highest to lowest precedence, alongside with their associativity.
The precedence and associativity of the operators is hard-coded into the parser.
Add parentheses if you want an expression with multiple operators to be interpreted in some other way than what the precedence rules imply.
| Operator | Description | Associativity |
|---|
. | Table/column name separator | left |
:: | Type cast | left |
[] | Array element | left |
+ - | Unary plus and minus | right |
AT | AT TIME ZONE | left |
^ | exponentiation | left |
* / % | multiplication, division, modulo | left |
+ / | addition, subtraction | left |
|| | concatenation | left |
BETWEEN | range containment | |
LIKE ILIKE | string matching | |
< > = <= >= <> | comparison | |
IS | IS NULL, IS DISTINCT FROM, IS TRUE | |
NOT | logical negation | right |
AND | logical conjunction | left |
OR | logical disjunction | left |
:: Type cast
Use can use the :: operator instead of the CAST function to convert one data type to another.
Syntax
-- CAST function
CAST(<value> AS <type>)
-- :: operator
<value>::<type>
| Component | Description |
|---|
<value> | The value to convert or an expression that results in a value to convert. Can be a column name, a function applied to a column or another function, or a literal value. |
<type> | The target data type (case-insensitive). |
Example
SELECT '2021-12-31'::DATE;
SELECT 8.5::REAL;
SELECT col_a::BIGINT;
Arithmetic with numbers
| Operator | Operator description | Input Data Types | Output Data Types | Example | Result |
|---|
| + | addition | INTEGER, BIGINT, NUMERIC, REAL, DOUBLE PRECISION | Corresponding type | SELECT 2 + 3; | 5 |
| - | subtraction | INTEGER, BIGINT, NUMERIC, REAL, DOUBLE PRECISION | Corresponding type | SELECT 2 - 3; | -1 |
| * | multiplication | INTEGER, BIGINT, NUMERIC, REAL, DOUBLE PRECISION | Corresponding type | SELECT 2 * 3; | 6 |
| / | division (integer division truncates the result) | INTEGER, BIGINT, NUMERIC, REAL, DOUBLE PRECISION | Corresponding type | SELECT 4 / 2; | 2 |
| % | modulo (remainder) | INTEGER, BIGINT | Corresponding type | SELECT 5 % 4; | 1 |
| ^ | exponentiation | INTEGER, BIGINT, NUMERIC, REAL, DOUBLE PRECISION | DOUBLE PRECISION | SELECT 2 ^ 3.0; | 8 |
In arithmetic operations like +, -, *, and / , the result’s data type aligns with the most encompassing type of the operands indicated as “Corresponding type” in the table above. For clarity:
- When both operands are of the same data type (e.g., two INTEGERs or two NUMERICs), the result will also be of that same data type.
INTEGER <op> INTEGER = INTEGERINTEGER <op> BIGINT = BIGINT
- For operations involving two different numeric data types, the result will typically be of the more precise or larger data type.
INTEGER <op> REAL = DOUBLE PRECISION
- Overflow checks and floating point errors are applied according to the result data type only.
Floating point precision means that the representation of a number is accurate up to a certain number of digits. In Firebolt, REAL data types have 6-digit precision and DOUBLE PRECISION have 16-digit precision. This means that calculations have a precision of 6 or 16 respectively, and numbers are truncated to that precision. For example, if a number is stored as 1.234567, it is automatically truncated to 1.23456 for REAL.
When performing arithmetic, the number of leading digits in the output is the product of the leading digits in both inputs. This means that if either or both of the input numbers are larger than 6, those numbers are the first truncated, and then the arithmetic is performed.
Arithmetic with date/time and intervals
An interval represents a duration.
In Firebolt, values of type interval can be used to add or subtract a duration to/from a date or timestamp.
Interval cannot be used as the data type of a column.
| Operator | Description |
|---|
DATE + interval -> TIMESTAMP | Add an interval to a DATE |
DATE - interval -> TIMESTAMP | Subtract an interval from a DATE |
TIMESTAMP + interval -> TIMESTAMP | Add an interval to a TIMESTAMP |
TIMESTAMP - interval -> TIMESTAMP | Subtract an interval from a TIMESTAMP |
TIMESTAMPTZ + interval -> TIMESTAMPTZ | Add an interval to a TIMESTAMPTZ |
TIMESTAMPTZ - interval -> TIMESTAMPTZ | Subtract an interval from a TIMESTAMPTZ |
interval * DOUBLE PRECISION -> interval | Multiply an interval by a scalar |
The + and * operators shown above come in commutative pairs (e.g., both DATE + interval and interval + DATE are accepted).
Literal string interpretation
Interval literals can be specified in two formats.
Cast from text literal
The following examples demonstrate different ways to cast a text literal to an interval in SQL. Each method converts a string representing a time interval into the INTERVAL data type using various casting syntaxes.
Using interval keyword:
interval 'quantity unit [quantity unit...] [direction]'
'quantity unit [quantity unit...] [direction]'::interval
CAST function:
CAST('quantity unit [quantity unit...] [direction]' AS interval)
direction can be either ago or left empty. Using ago negates all specified quantities. The quantity represents a signed or unsigned integer, and unit refers to one of the following time units, matched case-insensitively:
| Unit |
|---|
| microsecond[s] / us |
| millisecond[s] / ms |
| second[s] / s |
| minute[s] / m |
| hour[s] / h |
| day[s] / d |
| week[s] / w |
| month[s] / mon[s] |
| year[s] / y |
| decade[s] / dec[s] |
| century / centuries / c |
| millennium[s] / mil[s] |
Each unit can appear only once in an interval literal.
The value of the interval is determined by adding the quantities of the specified units with the appropriate signs.
Unit outside of text literal
The following example demonstrates how to cast a numeric value to an interval by placing the unit outside of the text literal.
In this format, N represents a signed or unsigned integer, and unit specifies the time unit, matched case-insensitively, from the following options:
| Unit |
|---|
| second |
| minute |
| hour |
| day |
| week |
| month |
| year |
Arithmetic between interval and TIMESTAMPTZ
Interval arithmetic with TIMESTAMPTZ values works as follows:
- Convert the
TIMESTAMPTZ value from Unix time to local time according to the rules of the time zone specified by the session’s time_zone setting.
- Add the
millennium, century, decade, year, month, week and day components of the interval to the local time.
- Convert the local time back to Unix time according to the rules of the time zone specified by the session’s
time_zone setting.
- Add the
hour, minute, second, millisecond, and microsecond components of the interval to the Unix time.
The back and forth between Unix time and local time is necessary to handle the fact that not all days consist of 24 hours due to daylight savings time transitions.
For instance, SELECT TIMESTAMPTZ '2022-10-30 Europe/Berlin' + interval '1 day' returns 2022-10-31 00:00:00+01 but SELECT TIMESTAMPTZ '2022-10-30 Europe/Berlin' + interval '24 hours' returns 2022-10-30 23:00:00+01 (assuming the value of the session’s time_zone setting is 'Europe/Berlin').
Still, the dependence on the session’s time_zone setting should be kept in mind when doing arithmetic between interval and TIMESTAMPTZ.
Multiplying an interval by a scalar
You can use the expression date_time + INTERVAL * d where date_time is a constant or column reference of type DATE, TIMESTAMP, or TIMESTAMPTZ, and d is a constant or column reference of type DOUBLE PRECISION.
The effect is that the INTERVAL is scaled by d, and the resulting INTERVAL is added to date_time.
E.g., INTERVAL '1 day' * 3 -> INTERVAL '3 days'.
Examples
SELECT DATE '1996-09-03' - interval '1 millennium 5 years 42 day 42 ms'; --> 0991-07-22 23:59:59.958
SELECT TIMESTAMP '1996-09-03 11:19:42' + interval '10 years 5 months 42 days 7 seconds'; --> 2007-03-17 11:19:49
SELECT TIMESTAMP '2023-10-20 11:49:52' + interval '1 year 6 months 4 weeks 7 hours' * 7.5; --> 2035-08-20 16:19:52
SELECT DATE '2023-10-20' - 42 * interval '1 months 1 day 1 hour'; --> 2020-03-07 06:00:00
-- The following example shows a daylight savings time change in the time zone 'Europe/Berlin'
SET time_zone = 'Europe/Berlin';
SELECT TIMESTAMPTZ '2022-10-30 Europe/Berlin' + interval '1 day'; --> 2022-10-31 00:00:00+01
SELECT TIMESTAMPTZ '2022-10-30 Europe/Berlin' + interval '24' hour; --> 2022-10-30 23:00:00+01
SET time_zone = 'US/Pacific';
SELECT TIMESTAMPTZ '2022-10-30 Europe/Berlin' + interval '1 day'; --> 2022-10-30 15:00:00-07
SELECT TIMESTAMPTZ '2022-10-30 Europe/Berlin' + interval '24' hour; --> 2022-10-30 15:00:00-07
Logical
Logical operators return the result of a boolean operation using three valued logic
| Operator | Example | Explanation |
|---|
AND | x AND y | True if both x and y are true |
NOT | NOT x | True if x is false |
OR | x OR y | True if either x or y is true |
Comparison
| Operator | Syntax | Explanation |
|---|
= | a=b | a is equal to b. |
!= | a!=b | a is not equal to b. |
<> | a<>b | a is not equal to b. |
<= | a<=b | a is less than or equal to b. |
> | a>b | a is greater than b. |
>= | a>=b | a is greater than or equal to b. |
< | a<b | a is less than b. |
BETWEEN | a BETWEEN b AND c | equivalent to b <= a <= c |
IS NULL | a IS NULL | a contains a NULL value |
IS NOT NULL | a IS NOT NULL | a doesn’t contain a NULL value |
IS NOT DISTINCT FROM | a IS NOT DISTINCT FROM b | equivalent to a=b where NULL is considered equal to NULL. |
IS DISTINCT FROM | a IS DISTINCT FROM b | equivalent to a!=b where NULL is considered equal to NULL. |
Example of using comparison operator in WHERE clause
SELECT
*
FROM
Table
WHERE
Price >= 100;
CASE
Conditional expression similar to if-then-else statements.
If the result of the condition is true, then the value of the CASE expression is the result that follows the condition. If the result is false, any subsequent WHEN clauses (conditions) are searched in the same manner. If no WHEN condition is true, then the value of the case expression is the result specified in the ELSE clause. If the ELSE clause is omitted and no condition matches, the result is NULL.
Syntax
CASE
WHEN <condition> THEN <result>
[ WHEN ...n ]
[ ELSE <result> ]
END;
Parameters
| Parameter | Description | Supported input types |
|---|
<condition> | A condition can be defined for each WHEN, and ELSE clause. | BOOLEAN |
<result> | The result of any condition. Every THEN clause receives a single result. All results in a single CASE function must share the same data type. | Any |
Return type
Same data type as <result>
Example
This example references a table player_level with the following columns and values:
| player | currentlevel |
|---|
| kennethpark | 3 |
| esimpson | 8 |
| sabrina21 | 11 |
| rileyjon | 15 |
| burchdenise | 4 |
The following example categorizes each entry by length. If the movie is longer than zero minutes and less than 50 minutes it is categorized as SHORT. When the length is 50-120 minutes, it’s categorized as Medium, and when even longer, it’s categorized as Long.
SELECT
player,
currentlevel,
CASE
WHEN length > 0
AND length <= 5 THEN 'Beginner'
WHEN length > 5
AND length <= 12 THEN 'Intermediate'
WHEN length > 12 THEN 'Expert'
END ranking
FROM
player_level
ORDER BY
player;
| player | currentlevel | ranking |
|---|
| kennethpark | 3 | Beginner |
| esimpson | 8 | Intermediate |
| sabrina21 | 11 | Intermediate |
| rileyjon | 15 | Expert |
| burchdenise | 4 | Beginner |
String
To concatenate strings, you can use the CONCAT function.
SELECT concat('This', ' is', ' a', ' parenthetical', 'concatenation.') AS concatenated_String
|| operator.
SELECT 'This' || ' is' || ' a' || ' double pipe' || ' concatenation.' AS concatenated_String
Subquery operators
Subqueries are queries nested within another query. They allow complex data retrieval by enabling a query to filter results based on the outcome of another query. Subquery operators are crucial in constructing these nested queries, especially within the WHERE clause, to filter data based on specific conditions.
| Operator | Explanation |
|---|
EXISTS | The EXISTS operator is used to check for the existence of any record in a subquery. It returns TRUE if the subquery returns one or more records. The subquery within EXISTS is executed repeatedly, once for each row that might be selected by the outer query. If the subquery returns any row, the EXISTS condition is met, and the outer query processing continues for that row. |
NOT EXISTS | It is opposite of EXISTS and is used to find records in one table that have no related records in another table. If the subquery returns no rows, NOT EXISTS returns TRUE. |
IN | The IN operator checks if a specific value is present in a list of values or the results of a subquery. Commonly utilized in a WHERE clause, it compares a column’s value against a predefined set. When the column’s value matches any value in this set, IN yields TRUE. |
NOT IN | Retrieve all entries from the value list that don’t match the required value. |
Example–using EXISTS to find all suppliers with products equal to the price of 22
SELECT supplier_name
FROM suppliers
WHERE EXISTS (
SELECT
product_name
FROM
products
WHERE
products.supplier_id = suppliers.supplier_id
AND
price = 22);
Example–using the IN operator to return all customers from Mannheim or London
SELECT
customer_name
FROM
customers
WHERE
customer_address IN ('Mannheim','London');
SELECT
product_id,
product_name,
list_price
FROM
products p
WHERE
list_price > (
SELECT
AVG( list_price )
FROM
products
WHERE
category_id = p.category_id);
Example–using a scalar boolean subquery to retrieve rows based on true/false condition
SELECT
*
FROM
products
WHERE (
SELECT CASE WHEN
MIN(list_price) > 100
THEN
true
ELSE
false
END
FROM
products);
