Selection Queries

On this page Carat arrow pointing down

Selection queries read and process data in CockroachDB. They are more general than simple SELECT clauses: they can group one or more selection clauses with set operations and can request a specific ordering or row limit.

Selection queries can occur:

Synopsis

Parameters

Parameter Description
common_table_expr See Common Table Expressions.
select_clause A valid selection clause, either simple or using set operations.
sort_clause An optional ORDER BY clause. See Ordering Query Results for details.
limit_clause An optional LIMIT clause. See Limiting Query Results for details.
offset_clause An optional OFFSET clause. See Limiting Query Results for details.
for_locking_clause The FOR UPDATE locking clause is used to order transactions by controlling concurrent access to one or more rows of a table. For more information, see SELECT FOR UPDATE.

The optional LIMIT and OFFSET clauses can appear in any order, but must appear after ORDER BY, if also present.

Note:
Because the WITH, ORDER BY, LIMIT and OFFSET sub-clauses are all optional, any simple selection clause is also a valid selection query.

Selection clauses

Selection clauses are the main component of a selection query. They define tabular data. There are four specific syntax forms collectively named selection clauses:

Form Usage
SELECT Load or compute tabular data from various sources. This is the most common selection clause.
VALUES List tabular data by the client.
TABLE Load tabular data from the database.
Set Operations Combine tabular data from two or more selection clauses.
Note:
To perform joins or other relational operations over selection clauses, use a table expression and convert it back into a selection clause with TABLE or SELECT.

Synopsis

VALUES clause

Syntax

VALUES ( a_expr , ) ,

A VALUES clause defines tabular data defined by the expressions listed within parentheses. Each parenthesis group defines a single row in the resulting table.

The columns of the resulting table data have automatically generated names. These names can be modified with AS when the VALUES clause is used as a sub-query.

Example

icon/buttons/copy
> VALUES (1, 2, 3), (4, 5, 6);
+---------+---------+---------+
| column1 | column2 | column3 |
+---------+---------+---------+
|       1 |       2 |       3 |
|       4 |       5 |       6 |
+---------+---------+---------+

TABLE clause

Syntax

A TABLE clause reads tabular data from a specified table. The columns of the resulting table data are named after the schema of the table.

In general, TABLE x is equivalent to SELECT * FROM x, but it is shorter to type.

Note:
Any table expression between parentheses is a valid operand for TABLE, not just simple table or view names.

Example

icon/buttons/copy
> CREATE TABLE employee_copy AS TABLE employee;

This statement copies the content from table employee into a new table. However, note that the TABLE clause does not preserve the indexing, foreign key, or constraint and default information from the schema of the table it reads from, so in this example, the new table employee_copy will likely have a simpler schema than employee.

Other examples:

icon/buttons/copy
> TABLE employee;
icon/buttons/copy
> INSERT INTO employee_copy TABLE employee;

SELECT clause

See Simple SELECT Clause for more details.

Set operations

Set operations combine data from two selection clauses. They are valid as operand to other set operations or as main component in a selection query.

Synopsis

select_clause UNION INTERSECT EXCEPT ALL DISTINCT select_clause

Set operators

SQL lets you compare the results of multiple selection clauses. You can think of each of the set operators as representing a Boolean operator:

  • UNION = OR
  • INTERSECT = AND
  • EXCEPT = NOT

By default, each of these comparisons displays only one copy of each value (similar to SELECT DISTINCT). However, each function also lets you add an ALL to the clause to display duplicate values.

Union: Combine two queries

UNION combines the results of two queries into one result.

icon/buttons/copy
> SELECT name
FROM accounts
WHERE state_opened IN ('AZ', 'NY')
UNION
SELECT name
FROM mortgages
WHERE state_opened IN ('AZ', 'NY');
+-----------------+
|      name       |
+-----------------+
| Naseem Joossens |
| Ricarda Caron   |
| Carola Dahl     |
| Aygün Sanna     |
+-----------------+

To show duplicate rows, you can use ALL.

icon/buttons/copy
> SELECT name
FROM accounts
WHERE state_opened IN ('AZ', 'NY')
UNION ALL
SELECT name
FROM mortgages
WHERE state_opened IN ('AZ', 'NY');
+-----------------+
|      name       |
+-----------------+
| Naseem Joossens |
| Ricarda Caron   |
| Carola Dahl     |
| Naseem Joossens |
| Aygün Sanna     |
| Carola Dahl     |
+-----------------+

Intersect: Retrieve intersection of two queries

INTERSECT finds only values that are present in both query operands.

icon/buttons/copy
> SELECT name
FROM accounts
WHERE state_opened IN ('NJ', 'VA')
INTERSECT
SELECT name
FROM mortgages;
+-----------------+
|      name       |
+-----------------+
| Danijel Whinery |
| Agar Archer     |
+-----------------+

Except: Exclude one query's results from another

EXCEPT finds values that are present in the first query operand but not the second.

icon/buttons/copy
> SELECT name
FROM mortgages
EXCEPT
SELECT name
FROM accounts;
+------------------+
|       name       |
+------------------+
| Günay García     |
| Karla Goddard    |
| Cybele Seaver    |
+------------------+

Ordering results

The following sections provide examples. For more details, see Ordering Query Results.

Order retrieved rows by one column

icon/buttons/copy
> SELECT *
FROM accounts
WHERE balance BETWEEN 350 AND 500
ORDER BY balance DESC;
+----+--------------------+---------+----------+--------------+
| id |        name        | balance |   type   | state_opened |
+----+--------------------+---------+----------+--------------+
| 12 | Raniya Žitnik      |     500 | savings  | CT           |
| 59 | Annibale Karga     |     500 | savings  | ND           |
| 27 | Adelbert Ventura   |     500 | checking | IA           |
| 86 | Theresa Slaski     |     500 | checking | WY           |
| 73 | Ruadh Draganov     |     500 | checking | TN           |
| 16 | Virginia Ruan      |     400 | checking | HI           |
| 43 | Tahirih Malinowski |     400 | checking | MS           |
| 50 | Dusan Mallory      |     350 | savings  | NV           |
+----+--------------------+---------+----------+--------------+

Order retrieved rows by multiple columns

Columns are sorted in the order you list them in sortby_list. For example, ORDER BY a, b sorts the rows by column a and then sorts rows with the same a value by their column b values.

icon/buttons/copy
> SELECT *
FROM accounts
WHERE balance BETWEEN 350 AND 500
ORDER BY balance DESC, name ASC;
+----+--------------------+---------+----------+--------------+
| id |        name        | balance |   type   | state_opened |
+----+--------------------+---------+----------+--------------+
| 27 | Adelbert Ventura   |     500 | checking | IA           |
| 59 | Annibale Karga     |     500 | savings  | ND           |
| 12 | Raniya Žitnik      |     500 | savings  | CT           |
| 73 | Ruadh Draganov     |     500 | checking | TN           |
| 86 | Theresa Slaski     |     500 | checking | WY           |
| 43 | Tahirih Malinowski |     400 | checking | MS           |
| 16 | Virginia Ruan      |     400 | checking | HI           |
| 50 | Dusan Mallory      |     350 | savings  | NV           |
+----+--------------------+---------+----------+--------------+

Limiting row count

You can reduce the number of results with LIMIT.

icon/buttons/copy
> SELECT id, name
FROM accounts
LIMIT 5;
+----+------------------+
| id |       name       |
+----+------------------+
|  1 | Bjorn Fairclough |
|  2 | Bjorn Fairclough |
|  3 | Arturo Nevin     |
|  4 | Arturo Nevin     |
|  5 | Naseem Joossens  |
+----+------------------+

Row-level locking for concurrency control with SELECT FOR UPDATE

The SELECT FOR UPDATE statement is used to order transactions by controlling concurrent access to one or more rows of a table.

It works by locking the rows returned by a selection query, such that other transactions trying to access those rows are forced to wait for the transaction that locked the rows to finish. These other transactions are effectively put into a queue based on when they tried to read the value of the locked rows.

Because this queueing happens during the read operation, the thrashing that would otherwise occur if multiple concurrently executing transactions attempt to SELECT the same data and then UPDATE the results of that selection is prevented. By preventing thrashing, CockroachDB also prevents transaction retries that would otherwise occur.

As a result, using SELECT FOR UPDATE leads to increased throughput and decreased tail latency for contended operations.

CockroachDB currently does not support the FOR SHARE/FOR KEY SHARE locking strengths, or the SKIP LOCKED wait policy.

Note:

By default, CockroachDB uses the SELECT FOR UPDATE locking mechanism during the initial row scan performed in UPDATE and UPSERT statement execution. To turn off implicit SELECT FOR UPDATE locking for UPDATE and UPSERT statements, set enable_implicit_select_for_update to false.

For an example showing how to use it, see SELECT FOR UPDATE.

Composability

Selection clauses are defined in the context of selection queries. Table expressions are defined in the context of the FROM sub-clause of SELECT. Nevertheless, they can be integrated with one another to form more complex queries or statements.

Using any selection clause as a selection query

Any selection clause can be used as a selection query with no change.

For example, the construct SELECT * FROM accounts is a selection clause. It is also a valid selection query, and thus can be used as a stand-alone statement by appending a semicolon:

icon/buttons/copy
> SELECT * FROM accounts;
+----+-----------------------+---------+----------+--------------+
| id |         name          | balance |   type   | state_opened |
+----+-----------------------+---------+----------+--------------+
|  1 | Bjorn Fairclough      |    1200 | checking | AL           |
|  2 | Bjorn Fairclough      |    2500 | savings  | AL           |
|  3 | Arturo Nevin          |     250 | checking | AK           |
[ truncated ]
+----+-----------------------+---------+----------+--------------+

Likewise, the construct VALUES (1), (2), (3) is also a selection clause and thus can also be used as a selection query on its own:

icon/buttons/copy
> VALUES (1), (2), (3);
+---------+
| column1 |
+---------+
|       1 |
|       2 |
|       3 |
+---------+
(3 rows)

Using any table expression as selection clause

Any table expression can be used as a selection clause (and thus also a selection query) by prefixing it with TABLE or by using it as an operand to SELECT * FROM.

For example, the simple table name customers is a table expression, which designates all rows in that table. The expressions TABLE accounts and SELECT * FROM accounts are valid selection clauses.

Likewise, the SQL join expression customers c JOIN orders o ON c.id = o.customer_id is a table expression. You can turn it into a valid selection clause, and thus a valid selection query as follows:

icon/buttons/copy
> TABLE (customers c JOIN orders o ON c.id = o.customer_id);
icon/buttons/copy
> SELECT * FROM customers c JOIN orders o ON c.id = o.customer_id;

Using any selection query as table expression

Any selection query (or selection clause) can be used as a table expression by enclosing it between parentheses, which forms a subquery.

For example, the following construct is a selection query, but is not a valid table expression:

icon/buttons/copy
> SELECT * FROM customers ORDER BY name LIMIT 5

To make it valid as operand to FROM or another table expression, you can enclose it between parentheses as follows:

icon/buttons/copy
> SELECT id FROM (SELECT * FROM customers ORDER BY name LIMIT 5);
icon/buttons/copy
> SELECT o.id
    FROM orders o
    JOIN (SELECT * FROM customers ORDER BY name LIMIT 5) AS c
      ON o.customer_id = c.id;

Using selection queries with other statements

Selection queries are also valid as operand in contexts that require tabular data.

For example:

Statement Example using SELECT Example using VALUES Example using TABLE
INSERT INSERT INTO foo SELECT * FROM bar INSERT INTO foo VALUES (1), (2), (3) INSERT INTO foo TABLE bar
UPSERT UPSERT INTO foo SELECT * FROM bar UPSERT INTO foo VALUES (1), (2), (3) UPSERT INTO foo TABLE bar
CREATE TABLE AS CREATE TABLE foo AS SELECT * FROM bar CREATE TABLE foo AS VALUES (1),(2),(3) CREATE TABLE foo AS TABLE bar
ALTER ... SPLIT AT ALTER TABLE foo SPLIT AT SELECT * FROM bar ALTER TABLE foo SPLIT AT VALUES (1),(2),(3) ALTER TABLE foo SPLIT AT TABLE bar
Subquery in a table expression SELECT * FROM (SELECT * FROM bar) SELECT * FROM (VALUES (1),(2),(3)) SELECT * FROM (TABLE bar)
Subquery in a scalar expression SELECT * FROM foo WHERE x IN (SELECT * FROM bar) SELECT * FROM foo WHERE x IN (VALUES (1),(2),(3)) SELECT * FROM foo WHERE x IN (TABLE bar)

See also


Yes No
On this page

Yes No