如何從 SQL 服務器中刪除轉換失敗的日期/時間錯誤？

Question

我寫了一個程序，它給了我這個錯誤

消息 241，級別 16，State 1，過程 SumByDay，第 7 行 [批處理開始第 36 行]
從字符串轉換日期和/或時間時轉換失敗。

這是程序

ALTER PROCEDURE SumByDay (@IN_forDate AS DATE, @IN_sensID INT)
AS
BEGIN
    DECLARE @timeFrom DATETIME2;
    DECLARE @timeTo DATETIME2;
    DECLARE @rowCount INT = 0; -- SQLINES DEMO *** statement affected rows

    SET @timeFrom = concat(CAST(@IN_forDate AS DATETIME2) , CAST('00:00:00' AS TIME)); -- start of the day
    SET @timeTo = concat(CAST(@IN_forDate AS DATETIME2) , CAST('23:59:59.999999' AS TIME)); -- end of the day
       
    DELETE FROM TrafficPerDay 
    WHERE TrafficPerDay.sensorID = @IN_sensID 
      AND TrafficPerDay.sDay = @IN_forDate; -- SQLINES DEMO *** sts, delete it, and recreate with new data
       
    INSERT INTO TrafficPerDay(sensorID, sDay, Motorcycle, Car, Van, Bus, Coach, Truck, ArticulatedTruck, Semitruck, Pedestrian, [Unknown], sTotal)
        (SELECT 
             t.sensorID, @IN_forDate AS 'sDay',
             COUNT(CASE WHEN t.classId = 1 THEN 'id' END) 'Motorcycle',
             COUNT(CASE WHEN t.classId = 2 THEN 'id' END) 'Car',
             COUNT(CASE WHEN t.classId = 3 THEN 'id' END) 'Van',
             COUNT(CASE WHEN t.classId = 4 THEN 'id' END) 'Bus',  
             COUNT(CASE WHEN t.classId = 5 THEN 'id' END) 'Coach',
             COUNT(CASE WHEN t.classId = 6 THEN 'id' END) 'Truck',
             COUNT(CASE WHEN t.classId = 7 THEN 'id' END) 'ArticulatedTruck',
             COUNT(CASE WHEN t.classId = 8 THEN 'id' END) 'Semitruck',
             COUNT(CASE WHEN t.classId = 9 THEN 'id' END) 'Pedestrian',
             COUNT(CASE WHEN t.classId = 0 THEN 'id' END) 'Unknown',
             COUNT(t.id) AS 'sTotal'
         FROM 
             TrafficRecord t
         WHERE 
             t.time BETWEEN @timeFrom AND @timeTo AND t.sensorId = @IN_sensID 
         GROUP BY t.sensorID, t.time );
        
    -- SQLINES LICENSE FOR EVALUATION USE ONLY
    SELECT @rowCount = @@ROWCOUNT

    IF (@rowCount = 0) 
    BEGIN  -- SQLINES DEMO *** ed, insert zeros
        INSERT INTO TrafficPerDay(sensorID, sDay, Motorcycle, Car, Van, Bus, Coach, Truck, ArticulatedTruck, Semitruck, Pedestrian, [Unknown], sTotal)
        VALUES (@IN_sensID, @IN_forDate, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
    END  
END

我正在執行這樣的程序

exec SumByDay @IN_forDate = '2020-11-04', @IN_sensID = 1;

Answer 1

問題是你在第 7 行的連接，正如錯誤告訴你的那樣，但你不需要那個。 使用>=和<邏輯而不是BETWEEN ：

ALTER PROCEDURE SumByDay (@IN_forDate AS date,
                          @IN_sensID int)
AS
BEGIN
    --DECLARE @timeFrom datetime2; -- No need for these
    --DECLARE @timeTo datetime2; -- No need for these
    DECLARE @rowCount int = 0; -- SQLINES DEMO *** statement affected rows
    --SET @timeFrom = CONCAT(CAST(@IN_forDate AS datetime2), CAST('00:00:00' AS time)); -- start of the day -- No need for these
    --SET @timeTo = CONCAT(CAST(@IN_forDate AS datetime2), CAST('23:59:59.999999' AS time)); -- end of the day -- No need for these

    DELETE FROM TrafficPerDay
    WHERE TrafficPerDay.sensorID = @IN_sensID
      AND TrafficPerDay.sDay = @IN_forDate; -- SQLINES DEMO ***
    --sts, delete it, and recreate with new data

    INSERT INTO TrafficPerDay (sensorID,sDay,Motorcycle,Car,Van,Bus,Coach,Truck,ArticulatedTruck,Semitruck,Pedestrian,[Unknown],sTotal)
    (SELECT t.sensorID,
            @IN_forDate AS sDay,
            COUNT(CASE WHEN t.classId = 1 THEN 'id' END) AS Motorcycle,
            COUNT(CASE WHEN t.classId = 2 THEN 'id' END) AS Car,
            COUNT(CASE WHEN t.classId = 3 THEN 'id' END) AS Van,
            COUNT(CASE WHEN t.classId = 4 THEN 'id' END) AS Bus,
            COUNT(CASE WHEN t.classId = 5 THEN 'id' END) AS Coach,
            COUNT(CASE WHEN t.classId = 6 THEN 'id' END) AS Truck,
            COUNT(CASE WHEN t.classId = 7 THEN 'id' END) AS ArticulatedTruck,
            COUNT(CASE WHEN t.classId = 8 THEN 'id' END) AS Semitruck,
            COUNT(CASE WHEN t.classId = 9 THEN 'id' END) AS Pedestrian,
            COUNT(CASE WHEN t.classId = 0 THEN 'id' END) AS Unknown,
            COUNT(t.id) AS sTotal
     FROM TrafficRecord t
     WHERE t.time >= @IN_forDate
       AND t.time < DATEADD(DAY, 1, @IN_forDate)
       AND t.sensorId = @IN_sensID
     GROUP BY t.sensorID,
              t.time); --Should this be CONVERT(date,t.time),
                       --or considering that this is for a single day, removed?
                       --If not, then simply t.time = @IN_forDate would work

    -- SQLINES LICENSE FOR EVALUATION USE ONLY
    SELECT @rowCount = @@ROWCOUNT;
    IF (@rowCount = 0)
    BEGIN -- SQLINES DEMO *** ed, insert zeros
        INSERT INTO TrafficPerDay (sensorID,sDay,Motorcycle,Car,Van,Bus,Coach,Truck,ArticulatedTruck,Semitruck,Pedestrian,[Unknown],sTotal)
        VALUES (@IN_sensID, @IN_forDate, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
    END;
END;

Answer 2

您設置的datetime2值不正確。

通過將時間連接到已經轉換為datetime2的日期上，您試圖將2021-03-18 00:00:00.000000000:00:00.0000000之類的日期分配給@timeFrom - 顯然不正確。

只需執行@timeFrom=Concat(@IN_forDate, ' 00:00:00')等