SELECT statement

Returns data from FeatureBase tables based on specified rows, columns and clauses.

Table of contents

BNF diagrams
Syntax
Arguments
Additional information
Examples
SELECT examples
Further information

BNF diagrams

expr

TOP clause

expr

SELECT list

expr expr

FROM clause

expr expr expr

WHERE clause

expr

GROUP BY clause

expr

HAVING clause

expr

ORDER BY clause

expr expr

Syntax

SELECT
  [DISTINCT]
  [<top_clause>]
  {<select_list>}
  {<from_clause>}
  [<where_clause>]
  [<group_by_clause>]
  [<having_clause>]
  [<order_by_clause>];

Arguments

Argument	Description	Required	Additional information
`DISTINCT`	Keyword that specifies only unique rows exist in the output	Optional	* DISTINCT additional * SELECT…flatten hint
top_clause	Specify a limit to apply to the number of rows returned in the output.	Optional	Requires integer literal
select_list	A series of expressions separated by commas that contains the items selected to form the output result set.	Yes	SELECT list and GROUP BY clause
from_clause	A list of table or subquery expressions that specify which relations to select data from.	Yes	FROM table or subquery expression
from…with	A list of table query hints	Optional for table queries	Query hints
where_clause	An expression that defines a filter condition for the rows returned by the query.	Optional	* `<expr>` filter conditions * SELECT…flatten hint
group_by_clause	Separates the results into groups of rows allowing aggregates to be performed on each group.	Optional	SELECT…flatten hint
column_expr	Specify a column or non-aggregate calculation on a column which is not required to appear in the select_list	Must exist in the from_clause
having_clause	Pass aggregates to filter on based on conditions.	Optional	`<expr>` filter conditions
order_by_clause	Comma-separated column name, column alias or column position in the SELECT list used to specify the order data is returned.	Optional	Ordering results

Additional information

DISTINCT additional

DISTINCT can return two kinds of results:

DISTINCT	Description	Examples
Values	A query that returns `DISTINCT` values from a table	SELECT from csv-target
Sets	A query that returns a specific array of values from `SET` or `SETQ` data type columns	Flatten hint

`<expr>` filter

<expr> can be used in the and clauses:

TOP
WHERE
HAVING

<expr> can include:

Filter	Description	Additional information
Subquery	A SELECT query that is run to obtain specific results required for the main query
Constant	A literal or scalar value	Can be joined by a subquery or SQL operators or subquery
Function	FeatureBase SQL functions	Can be joined by a subquery or SQL operators

SELECT list and GROUP BY clause

SELECT LIST Columns referenced in non-aggregate expressions must also appear in the GROUP BY clause.

SELECT…LIKE wildcards

Wildcards are used with the LIKE clause.

Wildcard	Description	Additional information
`*`	All columns
	Select List column alias
`expr`	Filter used to refine the query	`<expr>` filter
`<qualifier>.*`	limit the results to all columns based on the specified qualifier

SELECT with TUPLE() function

When the TUPLE() function is used in a select_list, the following values are returned:

Source data type	Returns
`IDSETQ`	`TIMESTAMP` and associated integer value
`STRINGSETQ`	`TIMESTAMP` and associated string value

Tuple() function

FROM table or subquery

The table_or_subquery expression can be:

a table_name or table_valued_function
a parenthesized SELECT statement

Both expressions can be aliased with a <table_alias>

Examples

`SELECT` examples

These examples demonstrate SELECT statements that use different clauses, expressions, functions and operators.

Statement	Target table	SELECT clauses	Functions	Operators
SELECT FROM cosvec-target	cosvec-target		COSINE DISTANCE
SELECT FROM csv-target	csv-target	* `COUNT()` * `DISTINCT`
SELECT FROM ndjson-target	parquet-target	* `WHERE` * `LIKE` * `ORDER BY`
SELECT FROM orc-target	orc-target
SELECT FROM parquet-target	parquet-target	* `TOP()` * `WHERE`		`BETWEEN`
SELECT FROM tan-target	tan-target		TANIMOTO_COEFFICIENT()
SELECT FROM tuple-set-target	parquet-target		DATETIMENAME()

SELECT int function

SELECT MIN(fld) FROM tbl
SELECT MAX(fld) FROM tbl
SELECT SUM(fld) FROM tbl
SELECT AVG(fld) FROM tbl
SELECT MIN(fld) FROM tbl WHERE fld = 1
SELECT MAX(fld) FROM tbl WHERE fld = 1
SELECT SUM(fld) FROM tbl WHERE fld = 1
SELECT AVG(fld) FROM tbl WHERE fld = 1

GROUP BY

SELECT fld, COUNT(*) FROM tbl GROUP BY fld

SELECT fld1, fld2, COUNT(*) FROM tbl GROUP BY fld1, fld2

SELECT fld1, fld2, COUNT(*) FROM tbl GROUP BY fld1, fld2 LIMIT 1

SELECT fld1, fld2, COUNT(*) FROM tbl WHERE fld1 = 1 GROUP BY fld1, fld2

SELECT fld1, COUNT(*) FROM tbl GROUP BY fld1 having COUNT(*) > 1

SELECT fld1, fld2, SUM(fld3) FROM tbl WHERE fld1 = 1 GROUP BY fld1, fld2

SELECT fld1, fld2, SUM(fld3) FROM tbl WHERE fld1 = 1 GROUP BY fld1, fld2 HAVING count(*) > 1

SELECT fld, COUNT(fld) FROM tbl GROUP BY fld

SELECT fld1, COUNT(fld1) FROM tbl WHERE fld2=1 GROUP BY fld1

GROUP BY with STRINGSET

This query that counts combinations of values from the segments table.

This query can also be run using the FLATTEN() hint

create table segments  
    (_id id, segment stringset);  

insert into segments(_id, segment)  
    values (1, ['RED', 'BLUE', 'GREEN']),
      (2, ['GREEN']),
      (3, ['RED', 'BLUE', 'GREEN']);

SELECT COUNT(*) as howmany, segment FROM segments
GROUP BY segment;

 howmany | segment
-----+--------------------------
   2 | ['RED', 'BLUE', 'GREEN']
   1 | ['GREEN']

This query can also be performed using the flatten hint