sql >> Database teknologi >  >> RDS >> Mysql

MySQL-forespørgsel til at beregne seneste gennemsnitspriser ved hjælp af data fra flere tabeller

En mulighed ville være en forespørgsel, der arbejder med brugervariabler for at generere en rangordning baseret på dato. Og vælg kun de tre sidste datoer.

Forespørgsel

SELECT 
 *
FROM (
   SELECT 
   * 
   , CASE 
       WHEN @supplier = supplier
       THEN @rank := @rank + 1 
       ELSE @rank := 1
     END 
      AS rank
   , @supplier := supplier 
   FROM 
     Expenses 
   CROSS JOIN (
    SELECT
       @supplier := NULL
     , @rank := 0
    ) 
    AS
      init_user_params  
   WHERE
      product_id = 1       
   ORDER BY
      supplier ASC 
    , DATE DESC   
)
 AS Expenses_ranked 
WHERE
  Expenses_ranked.rank <= 3

Resultat

    id  product_id  cost    quantity  supplier  date        @supplier := NULL  @rank := 0    rank  @supplier := supplier  
------  ----------  ------  --------  --------  ----------  -----------------  ----------  ------  -----------------------
     4           1  3.98           2         1  2017-09-22  (NULL)                      0       1                        1
     3           1  2.50           1         1  2017-09-20  (NULL)                      0       2                        1
     1           1  2.99           1         1  2017-09-05  (NULL)                      0       3                        1
     6           1  8.00           2         2  2017-09-27  (NULL)                      0       1                        2
     5           1  4.00           1         2  2017-09-25  (NULL)                      0       2                        2
     2           1  3.00           2         2  2017-09-10  (NULL)                      0       3                        2

Brug af disse resultater til at generere en gennemsnitsliste pr. leverandør.

Forespørgsel

SELECT 
   Expenses_ranked.supplier 
 , AVG(Expenses_ranked.cost / Expenses_ranked.quantity) AS AVG
FROM ( 

  SELECT 
   * 
   , CASE 
       WHEN @supplier = supplier
       THEN @rank := @rank + 1 
       ELSE @rank := 1
     END 
      AS rank
   , @supplier := supplier 
   FROM 
     Expenses 
   CROSS JOIN (
    SELECT
       @supplier := NULL
     , @rank := 0
    ) 
    AS
      init_user_params  
   WHERE
      product_id = 1       
   ORDER BY
      supplier ASC 
    , DATE DESC   
)
 AS Expenses_ranked 
WHERE
  Expenses_ranked.rank <= 3
GROUP BY
  Expenses_ranked.supplier 

Resultat

supplier  avg           
--------  --------------
       1  2.4933333333  
       2  3.1666666667  

Nu kan vi bruge en simpel ORDER BY [] ASC LIMIT 1 at få den billigste leverandør

Forespørgsel

SELECT 
 Expenses_ranked_avg.supplier AS cheapest_supplier
FROM ( 

  SELECT 
     Expenses_ranked.supplier 
   , AVG(Expenses_ranked.cost / Expenses_ranked.quantity) AS AVG
  FROM ( 

    SELECT 
     * 
     , CASE 
         WHEN @supplier = supplier
         THEN @rank := @rank + 1 
         ELSE @rank := 1
       END 
      AS rank
   , @supplier := supplier 
   FROM 
     Expenses 
   CROSS JOIN (
    SELECT
       @supplier := NULL
     , @rank := 0
    ) 
    AS
      init_user_params  
   WHERE
      product_id = 1       
   ORDER BY
      supplier ASC 
    , DATE DESC   
    )
      AS
        Expenses_ranked 
    WHERE
      Expenses_ranked.rank <= 3
    GROUP BY
      Expenses_ranked.supplier 
)
  AS Expenses_ranked_avg 
ORDER BY 
 Expenses_ranked_avg.avg ASC
LIMIT 1

Resultat

cheapest_supplier  
-------------------
                  1

Flere optimale forespørgsler.

Hvad er også muligt at erklære brugervariablerne i where-sætningen. Gør det direkte muligt at filtrere rangeringen fra.

Forespørgsel

  SELECT 
   *
  FROM
   Expenses 
  WHERE
   (
     CASE 
       WHEN @supplier = supplier
       THEN @rank := @rank + 1 
       ELSE @rank := 1
     END  
   ) 
 AND
   (@supplier := supplier )
 AND  
   @rank <= 3  
 AND  
   product_id = 1
ORDER BY 
   supplier ASC
 , DATE ASC  

Resultat

    id  product_id  cost    quantity  supplier  date        
------  ----------  ------  --------  --------  ------------
     1           1  2.99           1         1  2017-09-05  
     3           1  2.50           1         1  2017-09-20  
     4           1  3.98           2         1  2017-09-22  
     2           1  3.00           2         2  2017-09-10  
     5           1  4.00           1         2  2017-09-25  
     6           1  8.00           2         2  2017-09-27 

Nu er det nemt at bruge dette resultatsæt til at finde den billigste leverandør.

Forespørgsel

SELECT 
   Expenses_ranked.supplier AS cheapest_supplier
FROM (  

  SELECT 
   *
  FROM
   Expenses 
  WHERE
   (
     CASE 
       WHEN @supplier = supplier
       THEN @rank := @rank + 1 
       ELSE @rank := 1
     END  
   ) IS NOT NULL 
 AND
   (@supplier := supplier ) IS NOT NULL
 AND  
   @rank <= 3  
 AND  
   product_id = 1
ORDER BY 
   supplier ASC
 , DATE ASC  
)
 AS Expenses_ranked 
GROUP BY
  Expenses_ranked.supplier
ORDER BY 
  AVG(Expenses_ranked.cost / Expenses_ranked.quantity) ASC
LIMIT 1  

Resultat

cheapest_supplier  
-------------------
                  1


  1. Generering af Postgres-dump og lagring på en anden server

  2. Millisekunder i min DateTime ændres, når den er gemt i SQL Server

  3. Salesforce.com-linkede servere og sp_columns_ex

  4. arabiske tegn i oracle database