Skip to content

Latest commit

 

History

History
320 lines (242 loc) · 13.9 KB

expressions.rst

File metadata and controls

320 lines (242 loc) · 13.9 KB

Expressions

Table of contents

Expressions, particularly value expressions, are those which return a scalar value. Expressions have different types and forms. For example, there are literal values as atom expression and arithmetic, predicate and function expression built on top of them. And also expressions can be used in different clauses, such as using arithmetic expression in SELECT, WHERE or HAVING clause.

A literal is a symbol that represents a value. The most common literal values include:

  1. Numeric literals: specify numeric values such as integer and floating-point numbers.
  2. String literals: specify a string enclosed by single or double quotes.
  3. Boolean literals: true or false.
  4. Date and Time literals: DATE 'YYYY-MM-DD' represent the date, TIME 'hh:mm:ss' represent the time, TIMESTAMP 'YYYY-MM-DD hh:mm:ss' represent the timestamp. You can also surround the literals with curly brackets, if you do, you can replace date with d, time with t, and timestamp with ts

Here is an example for different type of literals:

os> SELECT 123, 'hello', false, -4.567, DATE '2020-07-07', TIME '01:01:01', TIMESTAMP '2020-07-07 01:01:01';
fetched rows / total rows = 1/1
+-----+---------+-------+--------+-------------------+-----------------+---------------------------------+
| 123 | 'hello' | false | -4.567 | DATE '2020-07-07' | TIME '01:01:01' | TIMESTAMP '2020-07-07 01:01:01' |
|-----+---------+-------+--------+-------------------+-----------------+---------------------------------|
| 123 | hello   | False | -4.567 | 2020-07-07        | 01:01:01        | 2020-07-07 01:01:01             |
+-----+---------+-------+--------+-------------------+-----------------+---------------------------------+


os> SELECT "Hello", 'Hello', "It""s", 'It''s', "It's", '"Its"', 'It\'s', 'It\\\'s', "\I\t\s"
fetched rows / total rows = 1/1
+---------+---------+---------+---------+--------+---------+---------+-----------+----------+
| "Hello" | 'Hello' | "It""s" | 'It''s' | "It's" | '"Its"' | 'It\'s' | 'It\\\'s' | "\I\t\s" |
|---------+---------+---------+---------+--------+---------+---------+-----------+----------|
| Hello   | Hello   | It"s    | It's    | It's   | "Its"   | It's    | It\'s     | \I\t\s   |
+---------+---------+---------+---------+--------+---------+---------+-----------+----------+


os> SELECT {DATE '2020-07-07'}, {D '2020-07-07'}, {TIME '01:01:01'}, {T '01:01:01'}, {TIMESTAMP '2020-07-07 01:01:01'}, {TS '2020-07-07 01:01:01'}
fetched rows / total rows = 1/1
+---------------------+------------------+-------------------+----------------+-----------------------------------+----------------------------+
| {DATE '2020-07-07'} | {D '2020-07-07'} | {TIME '01:01:01'} | {T '01:01:01'} | {TIMESTAMP '2020-07-07 01:01:01'} | {TS '2020-07-07 01:01:01'} |
|---------------------+------------------+-------------------+----------------+-----------------------------------+----------------------------|
| 2020-07-07          | 2020-07-07       | 01:01:01          | 01:01:01       | 2020-07-07 01:01:01               | 2020-07-07 01:01:01        |
+---------------------+------------------+-------------------+----------------+-----------------------------------+----------------------------+

The current implementation has the following limitations at the moment:

  1. Only literals of data types listed as above are supported for now. Other type of literals, such as NULL, will be added in future.
  2. Expression of literals, such as arithmetic expressions, will be supported later.
  3. Standard ANSI VALUES clause is not supported, although the SELECT literal example above is implemented by a Values operator internally.
  4. Date and Time literals only support DATE_FORMAT listed above.

Arithmetic expression is an expression formed by numeric literals and binary arithmetic operators as follows:

  1. +: Add.
  2. -: Subtract.
  3. *: Multiply.
  4. /: Divide. For integers, the result is an integer with fractional part discarded.
  5. %: Modulo. This can be used with integers only with remainder of the division as result.

Parentheses can be used to control the precedence of arithmetic operators. Otherwise, operators of higher precedence is performed first.

Implicit type conversion is performed when looking up operator signature. For example, an integer + a real number matches signature +(double,double) which results in a real number. This rule also applies to function call discussed below.

Here is an example for different type of arithmetic expressions:

os> SELECT 1 + 2, (9 - 1) % 3, 2 * 4 / 3;
fetched rows / total rows = 1/1
+-------+-------------+-----------+
| 1 + 2 | (9 - 1) % 3 | 2 * 4 / 3 |
|-------+-------------+-----------|
| 3     | 2           | 2         |
+-------+-------------+-----------+

Comparison operators are used to compare values. The MISSING and NULL value comparison has following the rule. MISSING value only equal to MISSING value and less than all the other values. NULL value equals to NULL value, large than MISSING value, but less than all the other values.

name description
> Greater than operator
>= Greater than or equal operator
< Less than operator
!= Not equal operator
<= Less than or equal operator
= Equal operator
LIKE Simple Pattern matching
IS NULL NULL value test
IS NOT NULL NOT NULL value test
IS MISSING MISSING value test
IS NOT MISSING NOT MISSING value test
REGEXP String matches regular expression test
IN IN value list test
NOT IN NOT IN value list test
BETWEEN Between a range (endpoint inclusive)
NOT BETWEEN Not between a range (BETWEEN negation)

Here is an example for different type of comparison operators:

os> SELECT 2 > 1, 2 >= 1, 2 < 1, 2 != 1, 2 <= 1, 2 = 1;
fetched rows / total rows = 1/1
+-------+--------+-------+--------+--------+-------+
| 2 > 1 | 2 >= 1 | 2 < 1 | 2 != 1 | 2 <= 1 | 2 = 1 |
|-------+--------+-------+--------+--------+-------|
| True  | True   | False | True   | False  | False |
+-------+--------+-------+--------+--------+-------+

It is possible to compare datetimes. When comparing different datetime types, for example DATE and TIME, both converted to TIMESTAMP. The following rule is applied on coversion: a TIME applied to today's date; DATE is interpreted at midnight. See example below:

os> SELECT current_time() > current_date() AS `now.time > today`, typeof(current_time()) AS `now.time.type`, typeof(current_date()) AS `now.date.type`;
fetched rows / total rows = 1/1
+------------------+---------------+---------------+
| now.time > today | now.time.type | now.date.type |
|------------------+---------------+---------------|
| True             | TIME          | DATE          |
+------------------+---------------+---------------+

os> SELECT current_time() = now() AS `now.time = now`, typeof(current_time()) AS `now.time.type`, typeof(now()) AS `now.type`;
fetched rows / total rows = 1/1
+----------------+---------------+-----------+
| now.time = now | now.time.type | now.type  |
|----------------+---------------+-----------|
| True           | TIME          | TIMESTAMP |
+----------------+---------------+-----------+

os> SELECT subtime(now(), current_time()) = current_date() AS `midnight = now.date`, typeof(subtime(now(), current_time())) AS `midnight.type`, typeof(current_date()) AS `now.date.type`;
fetched rows / total rows = 1/1
+---------------------+---------------+---------------+
| midnight = now.date | midnight.type | now.date.type |
|---------------------+---------------+---------------|
| True                | TIMESTAMP     | DATE          |
+---------------------+---------------+---------------+

expr LIKE pattern. The expr is string value, pattern is supports literal text, a percent ( % ) character for a wildcard, and an underscore ( _ ) character for a single character match, pattern is case insensitive:

os> SELECT 'axyzb' LIKE 'a%b', 'acb' LIKE 'A_B', 'axyzb' NOT LIKE 'a%b', 'acb' NOT LIKE 'a_b';
fetched rows / total rows = 1/1
+--------------------+------------------+------------------------+----------------------+
| 'axyzb' LIKE 'a%b' | 'acb' LIKE 'A_B' | 'axyzb' NOT LIKE 'a%b' | 'acb' NOT LIKE 'a_b' |
|--------------------+------------------+------------------------+----------------------|
| True               | True             | False                  | False                |
+--------------------+------------------+------------------------+----------------------+

Here is an example for null value test:

os> SELECT 0 IS NULL, 0 IS NOT NULL, NULL IS NULL, NULL IS NOT NULL;
fetched rows / total rows = 1/1
+-----------+---------------+--------------+------------------+
| 0 IS NULL | 0 IS NOT NULL | NULL IS NULL | NULL IS NOT NULL |
|-----------+---------------+--------------+------------------|
| False     | True          | True         | False            |
+-----------+---------------+--------------+------------------+

expr REGEXP pattern. The expr is string value, pattern is supports regular expression patterns:

os> SELECT 'Hello!' REGEXP '.*', 'a' REGEXP 'b';
fetched rows / total rows = 1/1
+----------------------+----------------+
| 'Hello!' REGEXP '.*' | 'a' REGEXP 'b' |
|----------------------+----------------|
| 1                    | 0              |
+----------------------+----------------+

Here is an example for IN value test:

os> SELECT 1 in (1, 2), 3 not in (1, 2);
fetched rows / total rows = 1/1
+-------------+-----------------+
| 1 in (1, 2) | 3 not in (1, 2) |
|-------------+-----------------|
| True        | True            |
+-------------+-----------------+

Here is an example for range test by BETWEEN expression:

os> SELECT
...  1 BETWEEN 1 AND 3,
...  4 BETWEEN 1 AND 3,
...  4 NOT BETWEEN 1 AND 3;
fetched rows / total rows = 1/1
+-------------------+-------------------+-----------------------+
| 1 BETWEEN 1 AND 3 | 4 BETWEEN 1 AND 3 | 4 NOT BETWEEN 1 AND 3 |
|-------------------+-------------------+-----------------------|
| True              | False             | True                  |
+-------------------+-------------------+-----------------------+

A function call is declared by function name followed by its arguments. The arguments are enclosed in parentheses and separated by comma. For complete function list supported, please see also: SQL Functions

A typical function call is in the following form:

function_name ( [ expression [, expression]* ]? )

If any argument is missing or null, the final result of evaluation will be missing or null accordingly.

Here is an example for different type of arithmetic expressions:

os> SELECT abs(-1.234), abs(-1 * abs(-5));
fetched rows / total rows = 1/1
+-------------+-------------------+
| abs(-1.234) | abs(-1 * abs(-5)) |
|-------------+-------------------|
| 1.234       | 5                 |
+-------------+-------------------+

Here is an example for different type of arithmetic expressions:

os> SELECT dayofmonth(DATE '2020-07-07');
fetched rows / total rows = 1/1
+-------------------------------+
| dayofmonth(DATE '2020-07-07') |
|-------------------------------|
| 7                             |
+-------------------------------+
  1. Only a subset of the SQL functions above is implemented in new engine for now. More function support are being added.
  2. For now function name is required to be lowercase.