为什么 MySQL 在 C# 代码中返回总是 1 而在我在存储过程中测试时不是？

Question

I am developing an ASP.net application using MySQL and have a question related to a stored procedure return value.

This is my stored procedure:

CREATE DEFINER=`pcg`@`%` PROCEDURE `UpdatePreSellerProfile`(
IN UserID INT(11),
IN SellerImageID INT(11),
IN BusinessImageID INT(11),
OUT ProfileUpdated INT(1)
)
BEGIN
SET @Approved = 'APPROVED';
UPDATE user SET 
    SELLER_IMAGE_ID = COALESCE((SELECT IMAGE_ID FROM image_url WHERE IMAGE_USER_ID = UserID AND IMAGE_ID=SellerImageID),SELLER_IMAGE_ID),
    SELLER_BUSINESS_LOGO_ID = COALESCE((SELECT IMAGE_ID FROM image_url WHERE IMAGE_USER_ID = UserID AND IMAGE_ID=BusinessImageID),SELLER_BUSINESS_LOGO_ID)
WHERE (USER_LOGIN_ID = UserID AND USER_PROFILE_STATUS = @Approved);
SET ProfileUpdated = ROW_COUNT();
END

When I test this code with following MySQL script I get 0 ( @ProfileUpdated ) always when there is no update.

call UpdatePreSellerProfile(@UserID, @SellerImageID, @BusinessImageID ,@ProfileUpdated);

But when I check this in my C# code, it is always showing 1 ( ProfileUpdated ).

if (oMySQLConnecion.State == System.Data.ConnectionState.Open)
{
    MySqlCommand oCommand = new MySqlCommand("UpdatePreSellerProfile", oMySQLConnecion);
    oCommand.CommandType = System.Data.CommandType.StoredProcedure;

    MySqlParameter sqlProfileUpdated = new MySqlParameter("@ProfileUpdated", MySqlDbType.VarString);
    sqlProfileUpdated.Direction = System.Data.ParameterDirection.Output;
    oCommand.Parameters.Add(sqlProfileUpdated);

    oCommand.Parameters.AddWithValue("@UserID", UserID);
    oCommand.Parameters.AddWithValue("@SellerImageID", oSeller.SellerImageID);
    oCommand.Parameters.AddWithValue("@BusinessImageID", oSeller.BusinessLogoID);

    oCommand.ExecuteNonQuery();
    Int16 ProfileUpdated = Convert.ToInt16(oCommand.Parameters["@ProfileUpdated"].Value);

    if (ProfileUpdated > 0) // <<-- Should be greater only if it is updated is sucessfull
    {
        oDBStatus.Type = DBOperation.SUCCESS;
        oDBStatus.Message.Add(DBMessageType.SUCCESSFULLY_DATA_UPDATED);
    }
    else
    {
        oDBStatus.Type = DBOperation.ERROR;
        oDBStatus.Message.Add(DBMessageType.ERROR_NO_RECORDS_UPDATED);
    }
    oMySQLConnecion.Close();
}

Why is the difference between MySQL script vs C# code?

Answer 1

Unless you have set the UseAffectedRows connection string option, it defaults to false . This means :

When false (default), the connection reports found rows instead of changed (affected) rows. Set to true to report only the number of rows actually changed by UPDATE or INSERT … ON DUPLICATE KEY UPDATE statements.

Additionally, from the documentation of the ROW_COUNT function :

For UPDATE statements, the affected-rows value by default is the number of rows actually changed. If you specify the CLIENT_FOUND_ROWS flag to mysql_real_connect() when connecting to mysqld [ed. note: this is the same as UseAffectedRows ], the affected-rows value is the number of rows “found”; that is, matched by the WHERE clause.

Thus, the UPDATE user statement in your stored procedure will return the number of rows that were found by the query, not the number that were actually updated.

To fix this, either:

1. Set UseAffectedRows=true; in your connection string; this may cause changes to other UPDATE queries.
2. Add more conditions to the WHERE clause, eg, WHERE ... AND SELLER_IMAGE_ID != SellerImageID AND SELLER_BUSINESS_LOGO_ID != BusinessImageID , to make sure the row is only found and updated if it actually needs to change.