Как вычислить значения в операторе обновления TSQL?

Question

Я хочу создать три временные таблицы, а затем объединить их содержимое так:

CREATE TABLE #TEMP1
MEMBERITEMCODE VARCHAR(25),
WEEK1USAGE VARCHAR(25),
WEEK1PRICE VARCHAR(25);

INSERT INTO #TEMP1 (MEMBERITEMCODE, WEEK1USAGE, WEEK1PRICE)
SELECT MEMBERITEMCODE, SUM(QTYSHIPPED), PRICE 
FROM INVOICEDETAIL 
WHERE UNIT=@UNIT AND INVOICEDATE BETWEEN @BEGDATE AND @WEEK1END
GROUP BY MEMBERITEMCODE, PRICE

CREATE TABLE #TEMP2
MEMBERITEMCODE VARCHAR(25),
WEEK2USAGE VARCHAR(25),
WEEK2PRICE VARCHAR(25);

INSERT INTO #TEMP2 (MEMBERITEMCODE, WEEK2USAGE, WEEK2PRICE)
SELECT MEMBERITEMCODE, SUM(QTYSHIPPED), PRICE 
FROM INVOICEDETAIL 
WHERE UNIT=@UNIT AND INVOICEDATE BETWEEN @WEEK2BEGIN AND @ENDDATE
GROUP BY MEMBERITEMCODE, PRICE

CREATE TABLE #TEMP3
MEMBERITEMCODE VARCHAR(25),
DESCRIPTION VARCHAR(200),
THIS VARCHAR(25),
THAT VARCHAR(25),
THEOTHERTHING VARCHAR(25);

INSERT INTO #TEMP3 (MEMBERITEMCODE, DESCRIPTION, THIS, THAT, THEOTHERTHING)
SELECT MEMBERITEMCODE, DESCRIPTION, THIS, THAT, THEOTHERTHING 
FROM INVOICEDETAIL 
WHERE UNIT=@UNIT AND INVOICEDATE BETWEEN @BEGDATE AND @ENDDATE

CREATE TABLE #TEMPCOMBINED
MEMBERITEMCODE VARCHAR(25),
DESCRIPTION VARCHAR(200),
THIS VARCHAR(25),
THAT VARCHAR(25),
THEOTHERTHING VARCHAR(25),
WEEK1USAGE VARCHAR(25),
WEEK1PRICE VARCHAR(25);
WEEK2USAGE VARCHAR(25),
WEEK2PRICE VARCHAR(25),
USAGEVARIANCE VARCHAR(25),
PRICEVARIANCE VARCHAR(25),
PRICEVARIANCEPERCENTAGE VARCHAR(25);

INSERT INTO #TEMPCOMBINED (MEMBERITEMCODE, DESCRIPTION, THIS, THAT, THEOTHERTHING, WEEK1USAGE, WEEK1PRICE, WEEK2USAGE, WEEK2PRICE, USAGEVARIANCE,
PRICEVARIANCE, PRICEVARIANCEPERCENTAGE)
SELECT T1.MEMBERITEMCODE, T3.DESCRIPTION, T3.THIS, T3.THAT, T3.THEOTHERTHING, T1.WEEK1USAGE, T1.WEEK1PRICE, T2.WEEK2USAGE, T2.WEEK2PRICE, NULL,  NULL,  NULL
FROM #TEMP1 T1
LEFT JOIN #TEMP2 T2 ON T1.MEMBERITEMCODE = T2.MEMBERITEMCODE
LEFT JOIN #TEMP3 T3 ON T1.MEMBERITEMCODE = T3.MEMBERITEMCODE

Теперь я хочу заменить пустые поля в вычисляемых полях, но не знаю, как это сделать. Моя лучшая идея на данный момент:

UPDATE #TEMPCOMBINED
SET USAGEVARIANCE = WEEK2USAGE - WEEK1USAGE,
PRICEVARIANCE = WEEK2PRICE - WEEK1PRICE,
PRICEVARIANCEPERCENTAGE = (WEEK2PRICE - WEEK1PRICE) / WEEK1PRICE

Есть ли лучший способ (например, тот, который действительно работает)?

Answer 1

Использование вычисляемых столбцов:

CREATE TABLE #TEMPCOMBINED(
  MEMBERITEMCODE VARCHAR(25),
  DESCRIPTION VARCHAR(200),
  THIS VARCHAR(25),
  THAT VARCHAR(25),
  THEOTHERTHING VARCHAR(25),
  WEEK1USAGE DECIMAL(18,10),
  WEEK1PRICE DECIMAL(18,10),
  WEEK2USAGE DECIMAL(18,10),
  WEEK2PRICE DECIMAL(18,10),
  USAGEVARIANCE   AS WEEK2USAGE - WEEK1USAGE,
  PRICEVARIANCE   AS WEEK2PRICE - WEEK1PRICE,
  PRICEVARIANCEPERCENTAGE  AS (WEEK2PRICE - WEEK1PRICE) / WEEK1PRICE
);

LiveDemo

И пропустить их в INSERT заявление:

INSERT INTO #TEMPCOMBINED (MEMBERITEMCODE, DESCRIPTION, THIS, THAT, THEOTHERTHING, WEEK1USAGE, WEEK1PRICE, WEEK2USAGE, WEEK2PRICE)
SELECT T1.MEMBERITEMCODE, T3.DESCRIPTION, T3.THIS, T3.THAT, T3.THEOTHERTHING, T1.WEEK1USAGE, T1.WEEK1PRICE, T2.WEEK2USAGE, T2.WEEK2PRICE
FROM #TEMP1 T1
LEFT JOIN #TEMP2 T2 ON T1.MEMBERITEMCODE = T2.MEMBERITEMCODE
LEFT JOIN #TEMP3 T3 ON T1.MEMBERITEMCODE = T3.MEMBERITEMCODE

Предупреждение:

Деление на WEEK1PRICE чтобы вы могли добавить (WEEK2PRICE - WEEK1PRICE) / NULLIF(WEEK1PRICE,0) чтобы избежать исключения Math или добавить CHECK ограничение для обеспечения того, чтобы WEEK1PRICE это не 0.

Answer 2

Первое: почему вы используете тип данных varchar для следующих столбцов?

USAGEVARIANCE, ,WEEK2USAGE,WEEK1USAGE,PRICEVARIANCE , WEEK2PRICE , WEEK1PRICE, PRICEVARIANCEPERCENTAGE, WEEK2PRICE , WEEK1PRICE,WEEK1PRICE ;

Лучше использовать десятичные или целочисленные типы данных. Как вы используете sum() функция для вычисления его значения, и он будет возвращать int/ десятичное значение

Второе: вы можете рассчитать USAGEVARIANCE,PRICEVARIANCE,PRICEVARIANCEPERCENTAGE вставляя в таблицу. Например:

select a,b, (a-b) as USAGEVARIANCE, (c-d) as PRICEVARIANCE from table 

и т.д., а затем вставьте его во временную таблицу

Answer 3

Обновление столбца в одной таблице позволяет использовать перекрестное соединение при простом объединении

Например:

Update m
Set col1 = col1 - col2
From mytable m 
join mytable n On m.col2 = n.col2