SnowConvert AI - Hive - SELECT¶

Applies to

Hive SQL
Spark SQL
Databricks SQL

Descrição¶

O Spark oferece suporte a uma instrução SELECT e está em conformidade com o padrão ANSI SQL. Consultas são usadas para recuperar conjuntos de resultados de uma ou mais tabelas. (Referência de linguagem SQL do Spark SELECT)

Aviso

Essa gramática é parcialmente suportada no Snowflake. Tradução pendente para estes elementos CREATE VIEW:

[ SORT BY { expression [ ASC | DESC ] [ NULLS { FIRST | LAST } ] [ , ... ] } ]
[ CLUSTER BY { expression [ , ... ] } ]
[ DISTRIBUTE BY { expression [, ... ] } ]
[ WINDOW { named_window [ , WINDOW named_window, ... ] } ]
[ PIVOT clause ]
[ UNPIVOT clause ]
[ LATERAL VIEW clause ] [ ... ]
[ regex_column_names ]
[ TRANSFORM (...) ]
[ LIMIT non_literal_expression ]

from_item :=
join_relation
table_value_function
LATERAL(subquery)
file_format.`file_path`

select_statement { INTERSECT | EXCEPT } { ALL | DISTINCT } select_statement

Sintaxe da gramática¶

[ WITH with_query [ , ... ] ]
select_statement [ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] select_statement, ... ]
    [ ORDER BY { expression [ ASC | DESC ] [ NULLS { FIRST | LAST } ] [ , ... ] } ]
    [ SORT BY { expression [ ASC | DESC ] [ NULLS { FIRST | LAST } ] [ , ... ] } ]
    [ CLUSTER BY { expression [ , ... ] } ]
    [ DISTRIBUTE BY { expression [, ... ] } ]
    [ WINDOW { named_window [ , WINDOW named_window, ... ] } ]
    [ LIMIT { ALL | expression } ]

select_statement :=
SELECT [ hints , ... ] [ ALL | DISTINCT ] { [ [ named_expression | regex_column_names ] [ , ... ] | TRANSFORM (...) ] }
    FROM { from_item [ , ... ] }
    [ PIVOT clause ]
    [ UNPIVOT clause ]
    [ LATERAL VIEW clause ] [ ... ] 
    [ WHERE boolean_expression ]
    [ GROUP BY expression [ , ... ] ]
    [ HAVING boolean_expression ]
    
with_query :=
expression_name [ ( column_name [ , ... ] ) ] [ AS ] ( query )

from_item :=
table_relation |
join_relation |
table_value_function |
inline_table |
LATERAL(subquery) |
file_format.`file_path` 

Amostra de padrões da origem¶

GROUP BY¶

A sintaxe WITH { CUBE | ROLLUP } é transformada para o equivalente CUBE(expr1, ...) ou ROLLUP(expr1, ...)

Código de entrada:¶

-- Basic case of GROUP BY
SELECT id, sum(quantity) FROM dealer GROUP BY 1;

-- Grouping by GROUPING SETS
SELECT city, car_model, sum(quantity) AS sum FROM dealer
    GROUP BY GROUPING SETS ((city, car_model), (city), (car_model), ());

-- Grouping by ROLLUP
SELECT city, car_model, sum(quantity) AS sum FROM dealer
    GROUP BY ROLLUP(city, car_model);

SELECT city, car_model, sum(quantity) AS sum FROM dealer
    GROUP BY city, car_model WITH ROLLUP;

-- Grouping by CUBE
SELECT city, car_model, sum(quantity) AS sum FROM dealer
    GROUP BY CUBE(city, car_model);

SELECT city, car_model, sum(quantity) AS sum FROM dealer
    GROUP BY city, car_model WITH CUBE;

Código de saída:¶

-- Basic case of GROUP BY
SELECT id,
    SUM(quantity) FROM
    dealer
GROUP BY 1;

-- Grouping by GROUPING SETS
SELECT city, car_model,
    SUM(quantity) AS sum FROM
    dealer
    GROUP BY GROUPING SETS ((city, car_model), (city), (car_model), () !!!RESOLVE EWI!!! /*** SSC-EWI-0073 - PENDING FUNCTIONAL EQUIVALENCE REVIEW FOR 'EmptyGroupingSet' NODE ***/!!!);

-- Grouping by ROLLUP
SELECT city, car_model,
    SUM(quantity) AS sum FROM
    dealer
    GROUP BY
    ROLLUP(city, car_model);

SELECT city, car_model,
    SUM(quantity) AS sum FROM
    dealer
GROUP BY
    ROLLUP(city, car_model);

-- Grouping by CUBE
SELECT city, car_model,
    SUM(quantity) AS sum FROM
    dealer
    GROUP BY CUBE(city, car_model) !!!RESOLVE EWI!!! /*** SSC-EWI-0073 - PENDING FUNCTIONAL EQUIVALENCE REVIEW FOR 'CUBE' NODE ***/!!!;

SELECT city, car_model,
    SUM(quantity) AS sum FROM
    dealer
GROUP BY
    CUBE(city, car_model);

Dicas¶

O Snowflake otimiza os JOINs e realiza o particionamento automaticamente, ou seja, dicas são desnecessárias, elas são preservadas como comentários no código de saída.

Código de entrada:¶

SELECT
/*+ REBALANCE */ /*+ COALESCE(2) */
*
FROM my_table;

Código de saída:¶

SELECT
/*+ REBALANCE */ /*+ COALESCE(2) */
*
FROM
my_table;

CTE¶

A palavra-chave AS é opcional no Spark/Databricks, mas no Snowflake é obrigatória, então ela é adicionada.

Código de entrada:¶

WITH my_cte (
   SELECT id, name FROM my_table
)
SELECT *
FROM my_cte
WHERE id = 1;

Código de saída:¶

WITH my_cte AS (
     SELECT id, name FROM
        my_table
  )
SELECT *
FROM
     my_cte
WHERE id = 1;

LIMIT¶

LIMIT ALL é removido porque não é necessário no Snowflake, o LIMIT com um valor literal é preservado como está.

Código de entrada:¶

SELECT * FROM my_table LIMIT ALL;

SELECT * FROM my_table LIMIT 5;

Código de saída:¶

SELECT * FROM
my_table;

SELECT * FROM
my_table
LIMIT 5;