​
Syntax

[ WITH clause ]
SELECT [ ALL | DISTINCT [ ON ( expression [, ...] ) ]] [ [table_name.]* [ EXCEPT ( [table_name.]except_column, ... ] ) ] | expression [ AS output_name ] [ , expression [ AS output_name ] ... ] ]
    [ VALUES clause ]
    [ FROM from_item [ , from_item ...] ]
    [ WHERE condition ]
    [ GROUP BY grouping_expression [ , grouping_expression ... ] ]
    [ HAVING condition ]
    [ ORDER BY sort_expression [ ASC | DESC ] [ NULLS { FIRST | LAST } ] [ , ... ] ]
    [ LIMIT count_number ]
    [ OFFSET start [ ROW | ROWS ] ];

Where from_item can be:

table_name  [ [ AS ] alias [ ( column_alias_list ) ] ]
    window_type ( table_name, col_name, interval_expression ) [ [ AS ] alias [ ( column_alias_list ) ] ]
    ( SELECT ) [ [ AS ] alias [ ( column_alias_list ) ] ]
    from_item join_type from_item [ ON join_condition ]

​
Parameters

Parameter or clauseDescription
WITH clauseProvides a way to write supplemental statements for a larger query. For more information, see WITH clause.
DISTINCT clauseThis clause eliminates duplicate rows from the result. SELECT DISTINCT eliminates duplicate rows based on all selected columns. SELECT DISTINCT ON allows you to specify expressions or columns and returns only the first row for each unique combination. It requires the use of the ORDER BY clause to determine the first row, and the DISTINCT ON expression must match the leftmost ORDER BY expression. The ORDER BY clause will normally contain additional expressions that determine the desired precedence of rows within each DISTINCT ON group. In this case, this expression can be an alternative with group topN when “N=1”. See examples of this clause below to know more about it.
EXCEPT clauseExclude one or more columns from the result set. By specifying except_column, the query will return all columns in the result set except those specified.
expressionA column or an expression.
VALUES clauseThis clause generates one or more rows of data as a table expression. For details, see VALUES clause.
aliasA temporary alternative name for a table or materialized view in a query.
table_nameA table or materialized view.
grouping_expressionValues can be:Input column namesInput column expressions without subqueries or correlated columns
ORDER BY clauseBy default, sorting is in ascending (ASC) order, with NULL values treated as the largest. For more information, see ORDER BY clause.
sort_expressionValues can be:Output column namesOutput column ordinal numbersHidden select expressions
count_numberThe number of results you want to get.
OFFSET clauseThe OFFSET clause can only be used with the LIMIT and ORDER BY clauses.
(SELECT)A SELECT command. You must enclose the subquery in parentheses, and specify an alias. When you include a subquery in the FROM clause, the output of the subquery is used as a temporary view that is only valid in the query.
join_typeSupported join types: [INNER] JOINLEFT [OUTER] JOINRIGHT [OUTER] JOINFULL [OUTER] JOINCurrently, only the ON clause is supported for joins.
join_conditionConditions for the ON clause that must be met before the two from_items can be joined.
window_typeThe type of the time window function. Possible values are HOP and TUMBLE.
interval_expressionThe interval expression, in the format of INTERVAL ‘<interval>’. For example: INTERVAL ‘2 MINUTES’. The standard SQL format, which places time units outside of quotation marks (for example, INTERVAL ‘2’ MINUTE), is also supported.
FROM clauseSpecifies the source of the data on which the query should operate. For more information, see FROM clause.
GROUP BY clauseGroups rows in a table with identical data, thus eliminating redundancy in the output and aggregates that apply to these groups. For more information, see GROUP BY clause.
HAVING clauseEliminates group rows that do not satisfy a given condition. For more information, see HAVING clause.
LIMIT clauseWhen the ORDER BY clause is not present, the LIMIT clause cannot be used as part of a materialized view. For more information, see LIMIT clause.
WHERE clauseSpecifies any conditions or filters to apply to your data. For more information, see WHERE clause.

​
Examples

​
DISTINCT clause

Here is an example of SELECT DISTINCT. This query will return only the unique combinations of first_name and last_name, eliminating any duplicate rows.

-- Retrieve the names of employees.
SELECT DISTINCT first_name, last_name
FROM employees;

Here is an example of SELECT DISTINCT ON. The query returns the latest order for each unique customer_id. ORDER BY is used to ensure that the desired row, that is, the row with the latest order date, appears first; otherwise, the returned row will be unpredictable.

-- Retrieve the latest order for each unique customer.
SELECT DISTINCT ON (customer_id) order_id, customer_id, order_date, total_amount
FROM orders
ORDER BY customer_id, order_date DESC;

​
Example of using several clauses

Below are the tables within the same schema that we will be writing queries from.

The table taxi_trips includes the columns id, distance, duration, and fare, where id identifies each unique trip.

{
  "id": VARCHAR,
  "distance": DOUBLE PRECISION,
  "duration": DOUBLE PRECISION,
  "fare": DOUBLE PRECISION
}

The table taxi includes the columns taxi_id and trip_id, where trip_id and id in taxi_trips are matching fields.

{
  "taxi_id": VARCHAR,
  "trip_id": VARCHAR
}

The table company includes the columns company_id and taxi_id, where taxi_id and taxi_id in taxi are matching fields.

{
  "company_id": VARCHAR,
  "taxi_id": VARCHAR
}

The following query returns the total distance and duration of trips that are beyond the initial charge ($2.50) of each taxi from the company “Yellow Taxi” and “FabCab”.

SELECT
    taxi.taxi_id,
    sum(trips.distance) AS total_distance,
    sum(trips.duration) AS total_duration
FROM taxi_trips AS trips
LEFT JOIN taxi ON trips.id = taxi.trip_id
WHERE taxi_id IN (
          SELECT taxi_id
          FROM company
          WHERE company_id IN ('Yellow Taxi', 'FabCab')
      )
      AND trips.fare > 2.50
GROUP BY taxi_id
ORDER BY total_distance, total_duration;