sql日期差异（以年为单位）-varchar列可能为空或null

我需要创建一个周年报告。这是原始代码

SELECT  cs.id AS csid ,
        csi.Search_Number AS searchID ,
        csi.date_of_placement AS datePlaced ,
        ui.firstname ,
        ui.lastname ,
        cp.name AS companyname ,
        cp.address_line_1 ,
        cp.address_line_2 ,
        cp.city ,
        cp.state ,
        cp.zip ,
        cp2.name AS currentcompanyname ,
        cp2.address_line_1 AS current_address_line_1 ,
        cp2.address_line_2 AS current_address_line_2 ,
        cp2.city AS current_city ,
        cp2.state AS current_state ,
        cp2.zip AS current_zip
FROM    client_searches_individuals AS csi
        LEFT JOIN users_info AS ui ON ui.id = csi.individual_number
        LEFT JOIN client_searches AS cs ON cs.id = csi.search_number
        LEFT JOIN companies AS Cp ON cp.id = cs.company_number
        LEFT JOIN companies AS cp2 ON cp2.id = ui.current_company_number
WHERE   0 = 0 
            <!--- This is for the varchar date field, take away anything with the year selected --->
                and csi.date_of_placement not like '#dateformat(attributes.startdate,'yyyy')#%'

            <cfif attributes.startdate neq "">
                and right(csi.date_of_placement,5) >= '#dateformat(attributes.startdate,'mm/dd')#'
            </cfif>
            <cfif attributes.enddate neq "">
                and right(csi.date_of_placement,5) <= '#dateformat(attributes.enddate,'mm/dd')#'
            </cfif>
        AND ISNULL(cs.id, 0) > 0
ORDER BY date_of_placement

我得到

味精242，第16级，第3状态，第2行 将char数据类型转换为datetime数据类型导致datetime值超出范围

SQL Server 2005 CF 9 我做了一个查询以更正放置日期。然后是第二个查询或第一个查询，以选择一年或更久的任何内容

SELECT  *
FROM    placedIndGoodDates
WHERE   0 = 0
        AND DATEDIFF(year, datePlaced, '#asofdate#') >= 1
ORDER BY date_of_placement

查询语法错误

遇到日期不同的年份。条件表达式不正确，应为[like | null | between | in | comparison]条件之一

---

我遗漏了一些基本信息。

考虑执行子选择，将放置日期设置为DATETIME，并确定放置日期为空或为空的记录的处理方法

您可以使用WHERE子句排除这些行，或者根据需要在将来为它们指定任意日期

一旦完成，您就可以使用日期范围比较而不产生问题，而不是手动比较年份等并遇到边界问题

更新：

根据额外信息，在将这些值输入DATEDIFF之前，使用WHERE子句删除无效条目，例如

SELECT
    CSI.OriginalDate,
    DATEDIFF(year, CSI.OriginalDate, GETDATE()) AS YearBoundariesCrossed,
    DATEDIFF(month, CSI.OriginalDate, GETDATE()) / 12 AS KindaFullYearsDifference
FROM (
    -- sub-select to remove invalid dates and convert to date
    SELECT
        CAST(date_of_placement AS datetime) AS OriginalDate,
        -- specify just the fields you need, or use the lazy *
        *
    FROM client_searches_individuals
    -- use whatever check(s) you need here, I like to check for LEN() > 0
    -- NULL entries fail this test too, which is good
    WHERE LEN(date_of_placement) > 0
) CSI
-- could join "CSI" to your other tables here, not doing that for this sample

-- normal date comparison OK here
WHERE CSI.OriginalDate < GETDATE()

对于额外的点，我已经包括了DATEDIFFyear之间的差异。。和DATEDIFFmonth，…/12如果您希望跨越至少12个月的边界，则后者更好

---

注：DATEDIFF只计算这一点，当使用“年”或“月”或“日”等时，跨越2013-12-31与2014-01-01的界限=1。

我不知道为什么，但使用dateadd代替DATEDIFF是有效的。你到底是如何使用dateadd的？因为它不能替代DateDiff.和csi.date_of_placement，它可以根据当前日期返回至少一年前的所有记录。但是，请注意，它依赖于隐式转换，这是不好的，原因有两个：一是您无法控制字符串的转换方式。结果可能并不总是您所期望的，并且可能根据db日期设置和表b中的实际值而有所不同。db将无法利用这些列上的索引。如果可能的话，最好将列数据类型更改为datetime。我希望重新设计数据库。但那是一种方式，一种方式，一种方式超出了范围。为什么日期存储为varchar？日期值应存储为日期，而不是字符串。这是一个遗留应用程序吗？

SELECT
    CSI.OriginalDate,
    DATEDIFF(year, CSI.OriginalDate, GETDATE()) AS YearBoundariesCrossed,
    DATEDIFF(month, CSI.OriginalDate, GETDATE()) / 12 AS KindaFullYearsDifference
FROM (
    -- sub-select to remove invalid dates and convert to date
    SELECT
        CAST(date_of_placement AS datetime) AS OriginalDate,
        -- specify just the fields you need, or use the lazy *
        *
    FROM client_searches_individuals
    -- use whatever check(s) you need here, I like to check for LEN() > 0
    -- NULL entries fail this test too, which is good
    WHERE LEN(date_of_placement) > 0
) CSI
-- could join "CSI" to your other tables here, not doing that for this sample

-- normal date comparison OK here
WHERE CSI.OriginalDate < GETDATE()