Question

我在SQL中有307807行的表。如果我的start_date和end_date列具有无效值，我想删除记录。不幸的是，我的userID重复了不止一次。因此，我想删除受无效值影响的用户ID。

UserID ---- start_date ---- end_date
123    ---- 06-05-1934 ---- 16-05-1934
123    ---- 06-05-1933 ---- 16-05-1933
123    ---- /N         ---- 16-03-1920
234    ---- 04-03-1945 ---- 17-04-1945
234    ---- 05-06-1999 ---- 06-07-1999

这里我想删除所有带有用户ID 123的记录，因为我不能再使用该数据。

我使用MS SQL Server Management Studio 17。

Answer 1

使用;WITH cte AS (SELECT indictr = Min(CASE WHEN Try_convert(datetime, start_date, 105) IS NULL OR Try_convert(datetime, end_date, 105) IS NULL THEN 0 ELSE 1 END) OVER(partition BY UserID ),* FROM Yourtable) DELETE FROM cte WHERE indictr = 0检查token_url = "https://api.fitbit.com/oauth2/token" access_url = "https://api.fitbit.com/oauth2/token" auth_url = "https://www.fitbit.com/oauth2/authorize" key = "XXXX" secret = "XXXX" fbr = oauth_app("Glinski",key,secret) fitbit = oauth_endpoint(token_url,auth_url,access_url) token = oauth2.0_token(fitbit,fbr, scope = c("activity","heartrate"))是否有效

GET(url = "http://api.fitbit.com/1/user/-/activities/heartrate/date/2018-01-
01/today.json", config(token = token))

Answer 2

您需要选择日期不佳的所有用户，然后删除这些用户的记录：

WITH BadUsers as (
SELECT DISTINCT UserID from TableA 
where start_date = '/n' or end_date ='n'
)
DELETE FROM TableA
JOIN BadUsers ON
Table.UserID = BadUser.UserID

Answer 3

您可以使用 TRY_CAST (Transact-SQL) 作为：

CREATE TABLE T (
    ID INT,
    SDate VARCHAR(10),
    EDate VARCHAR(10)
    );

INSERT INTO T VALUES
(123, '1934-05-06', '1634-05-16'),
(123, '/N', '1934-05-25'),
(123, '1934-05-26', 'WrongDate');

SELECT *
FROM T;

DELETE FROM T
WHERE ID IN 
        (
         SELECT DISTINCT ID 
         WHERE (TRY_CAST(SDate AS DATE) IS NULL) OR (TRY_CAST(EDate AS DATE) IS NULL) 
        );

SELECT *
FROM T;

如果一条记录的值无效，则删除具有相同ID的记录

3 个答案: