Question

我要做的是：我创建了一个T-SQL存储过程，它从多个表中返回一个列。

CREATE PROCEDURE [dbo].[Search_] 
    (@am VARCHAR(12)) 
AS
BEGIN
    SET NOCOUNT ON;

    DECLARE @DBName VARCHAR(128)

    DECLARE Tbl CURSOR READ_ONLY FOR
         SELECT LEFT(TABLE_NAME, 27) AS Tbl 
         FROM ap.INFORMATION_SCHEMA.TABLES 
         WHERE TABLE_NAME LIKE LEFT('STORED_PRODUCTS_FR_[0-9]_%', 27)
           AND SUBSTRING(TABLE_NAME, 20, 2) LIKE '%[0-9]%'
           AND SUBSTRING(TABLE_NAME, 23, 1) = '_'
         ORDER BY
             CONVERT(INT, SUBSTRING(TABLE_NAME, 24, 4)) ASC,
             CONVERT(INT, SUBSTRING(TABLE_NAME, 21, 2)) ASC

    OPEN Tbl

    FETCH NEXT FROM Tbl INTO @DBName

    WHILE @@FETCH_STATUS = 0
    BEGIN
        DECLARE @SQL VARCHAR(100)
        SET @sql = 'SELECT [column 3] FROM ' + @DBName + ' WHERE [COLUMN 4] = ' + @am + '';

        EXEC(@sql)

        FETCH NEXT FROM Tbl INTO @DBName
    END
END

CLOSE Tbl
DEALLOCATE Tbl

然后在asp.net中，我想填充gridview ...

try
{
    SqlDataAdapter dt = new SqlDataAdapter();

    string message = string.Empty;
    con.Open();

    cmd.CommandText = "Search_";
    cmd.CommandType = CommandType.StoredProcedure;
    cmd.Parameters.AddWithValue("@am", am.Text);   
    cmd.Connection = con;

    dt.SelectCommand = cmd;
    DataTable dTable = new DataTable();
    dt.Fill(dTable);

    GridView1.DataSource = dTable;
    GridView1.DataBind();
}
catch (Exception ex)
{
    throw ex;
}

只返回存储过程的单行（第一行）。

我做错了什么？似乎价值会覆盖自己。

任何人的帮助......

Answer 1

您的存储过程返回两行，但由于在另一个select语句中选择了每一行，DataAdapter只能使用返回的第一行来填充DataTable。如果您填写DataSet，则会包含2个DataTable个实例。

请注意您的存储过程表明您的数据库设计存在缺陷 - 拥有多个存储相同类型实体的表对数据库来说是一个糟糕的设计。但是，我不确定这是不是。

无论如何，要使用快速补丁解决此问题，您可以将过程更改为以下内容：

CREATE PROCEDURE [dbo].[Search_] (@am varchar(12)) 
AS
BEGIN
    SET NOCOUNT ON;
    DECLARE @DBName varchar(128)
    declare @SQL varchar(max) = ''

    DECLARE Tbl CURSOR READ_ONLY FOR
    select LEFT(TABLE_NAME,27) as Tbl from ap.INFORMATION_SCHEMA.TABLES 
    where TABLE_NAME like left('STORED_PRODUCTS_FR_[0-9]_%',27)
    and SUBSTRING(TABLE_NAME , 20,2) like '%[0-9]%'
    and SUBSTRING(TABLE_NAME , 23,1) = '_'
    order by CONVERT(int , substring(TABLE_NAME,24,4)) asc ,
    CONVERT(int , substring(TABLE_NAME,21,2)) asc   

    OPEN Tbl

    FETCH NEXT FROM Tbl INTO @DBName
    WHILE @@FETCH_STATUS = 0
    BEGIN
        set @sql += 'union all select [column 3] from '+@DBName+' WHERE [COLUMN 4] ='+@am+' ';
        FETCH NEXT FROM Tbl INTO @DBName
    END

    CLOSE Tbl
    DEALLOCATE Tbl

    SET @sql = STUFF(@Sql, 1, 10, '') -- Remove the first union all
    exec(@sql)
END

Answer 2

在你的脚本中，最后end应放在DEALLOCATE Tbl之后，并且为了调用sp，它会更好地返回多个记录集以使用DataSet或DataReader：

var dataset = new DataSet();
using(var cnnn = new SqlConnection(cnnString))
{
    var adapter = new SqlDataAdapter();
    adapter.SelectCommand = new SqlCommand("dbo.Search_", cnnn);
    adapter.SelectCommand.CommandType = CommandType.StoredProcedure;
    adapter.Fill(dataset);
}
dataset.Tables => gives u all the results of exec(@sql)

因为每个结果集都返回相同的列名和结构，所以你可以将它们合并为一个tb：

var allTbs = dataset.Tables[0].Copy();
for(int i = 1; i < dataset.Tables.Count; i++)
    allTbs.Merge(dataset.Tables[i]);

tsql过程返回值asp.net

2 个答案: