Sql 批量插入格式文件中的正确排序规则

我正在尝试使用带格式文件的大容量插入将.CSV文件导入SQL Server表。我可以让它导入，但任何拉丁字符都是作为奇怪字符导入的。我为自己完成这个个人项目感到非常自豪，但我已经到了需要帮助的地步。导入数据后，我可以通过执行一些凌乱的UPDATE和REPLACE语句来更改字符，但我确实希望能够一步导入.CSV文件中显示的拉丁字符。 以下是我创建的数据库和表：

CREATE DATABASE Test;

CREATE TABLE dbo.rawData
    ([Position] nvarchar(500) NULL,
    [Const] nvarchar(500) NULL,
    [Created] nvarchar(500) NULL,
    [Modified] nvarchar(500) NULL,
    [Description] nvarchar(500) NULL,
    [Title] nvarchar(500) NOT NULL,
    [TitleType] nvarchar(500) NULL,
    [Directors] nvarchar(500) NULL,
    [YouRated] nvarchar(500) NULL,
    [IMDbRating] nvarchar(500) NULL,
    [Runtime] nvarchar(500) NULL,
    [Year] nvarchar(500) NULL,
    [Genres] nvarchar(500) NULL,
    [NumVotes] nvarchar(500) NULL,
    [ReleaseDate] nvarchar(500) NULL, 
    [URL] nvarchar(500) NULL,
    )
GO

下面是我正在处理的一些数据，它们来自.CSV文件（另存为ratings.CSV）。我使用记事本+，它是用UTF-8编码的。请注意，“达拉斯买家俱乐部”的最后一行是如何有一个名字中带有拉丁字符的董事的：

"position","const","created","modified","description","Title","Title type","Directors","You rated","IMDb Rating","Runtime (mins)","Year","Genres","Num. Votes","Release Date (month/day/year)","URL"
"1","tt0437863","Tue Feb 16 00:00:00 2016","","","The Benchwarmers","Feature Film","Dennis Dugan","5","5.6","80","2006","comedy, romance, sport","39413","2006-04-07","http://www.imdb.com/title/tt0437863/"
"2","tt0085334","Tue Feb 16 00:00:00 2016","","","A Christmas Story","Feature Film","Bob Clark","6","8.1","94","1983","comedy, family","103770","1983-11-18","http://www.imdb.com/title/tt0085334/"
"3","tt2403029","Tue Feb 16 00:00:00 2016","","","The Starving Games","Feature Film","Jason Friedberg, Aaron Seltzer","2","3.3","83","2013","comedy","13719","2013-10-31","http://www.imdb.com/title/tt2403029/"
"4","tt0316465","Tue Feb 16 00:00:00 2016","","","Radio","Feature Film","Michael Tollin","6","6.9","109","2003","biography, drama, sport","31692","2003-10-24","http://www.imdb.com/title/tt0316465/"
"5","tt0141369","Tue Feb 16 00:00:00 2016","","","Inspector Gadget","Feature Film","David Kellogg","4","4.1","78","1999","action, adventure, comedy, family, sci_fi","35340","1999-07-18","http://www.imdb.com/title/tt0141369/"
"6","tt0033563","Tue Feb 16 00:00:00 2016","","","Dumbo","Feature Film","Sam Armstrong, Norman Ferguson","6","7.3","64","1941","animation, family, musical","80737","1941-10-23","http://www.imdb.com/title/tt0033563/"
"7","tt0384642","Tue Feb 16 00:00:00 2016","","","Kicking & Screaming","Feature Film","Jesse Dylan","5","5.5","95","2005","comedy, family, romance, sport","29539","2005-05-01","http://www.imdb.com/title/tt0384642/"
"8","tt0116705","Tue Feb 16 00:00:00 2016","","","Jingle All the Way","Feature Film","Brian Levant","7","5.4","89","1996","comedy, family","66879","1996-11-16","http://www.imdb.com/title/tt0116705/"
"9","tt1981677","Tue Feb 16 00:00:00 2016","","","Pitch Perfect","Feature Film","Jason Moore","7","7.2","112","2012","comedy, music, romance","203205","2012-09-28","http://www.imdb.com/title/tt1981677/"
"10","tt0409459","Tue Feb 16 00:00:00 2016","","","Watchmen","Feature Film","Zack Snyder","7","7.6","162","2009","action, mystery, sci_fi","368137","2009-02-23","http://www.imdb.com/title/tt0409459/"
"11","tt1343092","Tue Feb 16 00:00:00 2016","","","The Great Gatsby","Feature Film","Baz Luhrmann","5","7.3","143","2013","drama, romance","345664","2013-05-01","http://www.imdb.com/title/tt1343092/"
"12","tt0332379","Tue Feb 16 00:00:00 2016","","","School of Rock","Feature Film","Richard Linklater","5","7.1","108","2003","comedy, music","202083","2003-09-09","http://www.imdb.com/title/tt0332379/"
"13","tt0120783","Tue Feb 16 00:00:00 2016","","","The Parent Trap","Feature Film","Nancy Meyers","6","6.4","128","1998","adventure, comedy, drama, family, romance","82087","1998-07-20","http://www.imdb.com/title/tt0120783/"
"14","tt0790636","Tue Feb 16 00:00:00 2016","","","Dallas Buyers Club","Feature Film","Jean-Marc Vallée","7","8.0","117","2013","biography, drama","308118","2013-09-07","http://www.imdb.com/title/tt0790636/"

我有一个格式文件（另存为format.fmt），在Notepad++中打开时如下所示：

11.0
16
1       SQLCHAR             0       1000    "\",\""    1     Position                   SQL_Latin1_General_CP1_CI_AS
2       SQLCHAR             0       1000    "\",\""    2     Const                      SQL_Latin1_General_CP1_CI_AS
3       SQLCHAR             0       1000    "\",\""    3     Created                    SQL_Latin1_General_CP1_CI_AS
4       SQLCHAR             0       1000    "\",\""    4     Modified                   SQL_Latin1_General_CP1_CI_AS
5       SQLCHAR             0       1000    "\",\""    5     Description                SQL_Latin1_General_CP1_CI_AS
6       SQLCHAR             0       1000    "\",\""    6     Title                      SQL_Latin1_General_CP1_CI_AS
7       SQLCHAR             0       1000    "\",\""    7     TitleType                  SQL_Latin1_General_CP1_CI_AS
8       SQLCHAR             0       1000    "\",\""    8     Directors                  SQL_Latin1_General_CP1_CI_AS
9       SQLCHAR             0       1000    "\",\""    9     YouRated                   SQL_Latin1_General_CP1_CI_AS
10      SQLCHAR             0       1000    "\",\""    10    IMDbRating                 SQL_Latin1_General_CP1_CI_AS
11      SQLCHAR             0       1000    "\",\""    11    Runtime                    SQL_Latin1_General_CP1_CI_AS
12      SQLCHAR             0       1000    "\",\""    12    Year                       SQL_Latin1_General_CP1_CI_AS
13      SQLCHAR             0       1000    "\",\""    13    Genres                     SQL_Latin1_General_CP1_CI_AS
14      SQLCHAR             0       1000    "\",\""    14    NumVotes                   SQL_Latin1_General_CP1_CI_AS
15      SQLCHAR             0       1000    "\",\""    15    ReleaseDate                SQL_Latin1_General_CP1_CI_AS
16      SQLCHAR             0       1000    "\""     16    URL                        SQL_Latin1_General_CP1_CI_AS

当我运行下面的代码时，所有内容都导入，但是拉丁字符被一系列奇怪的字符替换。以下是我正在运行的代码：

BULK INSERT [Test].[dbo].[rawData]
FROM 'C:\IMDbRatings\Files\ratings.csv' WITH (FIRSTROW = 2, FORMATFILE= 'C:\IMDbRatings\format.fmt');

---

我做了一些尝试，将.CSV文件更改为UCS-2 BE，在大容量插入的WITH子句中添加了不同的条件，并将格式文件中的变量类型更改为SQLNCHAR而不是SQLCHAR，但没有任何效果。在这些情况下，通常发生的是“0行受影响”，而不是错误。任何帮助都将不胜感激。

@Walker我承认我从未使用过大容量插入，但试图设置您的测试用例，只是不断变得不完整或无法读取我已保存的格式文件。无论如何，请尝试在编码-->字符集-->Western Eurpoean-->Windows-1252的记事本++中将编码更改为1252保存文件并尝试导入

---

另外，我刚刚看到了这篇有趣的文章，它认为UTF-8在SQL 2016之前是一个问题。但是吸引我注意的一个答案是

SQLNCHAR vs SQLCHAR

，因为我认为您存储的是

Unicode

数据，这意味着您需要更改已装箱的格式文件和表中的数据类型。

我回答这个老问题，希望它能让人省去我最近遇到的麻烦通过

简单地说：当使用代码页65001从UTF-8编码文件插入时，您应该在格式文件中使用

”

排序规则。您必须拥有SQL Server 2016，才能使用代码页65001

---

请执行以下操作：

在bulk insert语句中指定大容量插入表以UTF-8编码，并使用

CODEPAGE=65001

在格式化文件中，将字符列类型指定为SQLCHAR

在格式化文件中，对所有列使用“”排序规则

批量插入语句：

BULK INSERT [Test].[dbo].[rawData]
FROM 'C:\IMDbRatings\Files\ratings.csv'
WITH (CODEPAGE = 65001, FIRSTROW = 2, FORMATFILE= 'C:\IMDbRatings\format.fmt');

格式文件：

13.0
16
1       SQLCHAR             0       1000    "\",\""    1     Position                   ""
2       SQLCHAR             0       1000    "\",\""    2     Const                      ""
3       SQLCHAR             0       1000    "\",\""    3     Created                    ""
4       SQLCHAR             0       1000    "\",\""    4     Modified                   ""
5       SQLCHAR             0       1000    "\",\""    5     Description                ""
6       SQLCHAR             0       1000    "\",\""    6     Title                      ""
7       SQLCHAR             0       1000    "\",\""    7     TitleType                  ""
8       SQLCHAR             0       1000    "\",\""    8     Directors                  ""
9       SQLCHAR             0       1000    "\",\""    9     YouRated                   ""
10      SQLCHAR             0       1000    "\",\""    10    IMDbRating                 ""
11      SQLCHAR             0       1000    "\",\""    11    Runtime                    ""
12      SQLCHAR             0       1000    "\",\""    12    Year                       ""
13      SQLCHAR             0       1000    "\",\""    13    Genres                     ""
14      SQLCHAR             0       1000    "\",\""    14    NumVotes                   ""
15      SQLCHAR             0       1000    "\",\""    15    ReleaseDate                ""
16      SQLCHAR             0       1000    "\""     16    URL                        ""

---

在

”

或

RAW

排序规则中：

指定数据存储在指定的代码页中 在命令或bcp_控件BCPFILECP的代码页选项中 暗示如果未指定任何一项，则数据文件的排序规则为 客户端计算机的OEM代码页

---

您的数据库、表和列的“Jean-Marc Vallée”名称是什么样的排序规则？这三个名称都使用SQL\u Latin\u General\u CP1\u CI\u，因为您几天前没有发布完全相同的问题吗？导入方法是否必须使用bulkinsert？您是否有可用的SSI？是否尝试在代码页1252的记事本++中保存文件并导入？这建议在Bulkinsert语句中为格式文件或rating.csv使用CODEPAGE='ACP'？我真的很感谢你在这里的帮助。rating.csv我不认为格式文件会有什么不同。我还写了一点。我认为你需要将数据类型修改为nchar，这样你就可以处理Unicode字符，因为char不会存储它们。这当然很奇怪，但编码方面肯定会发生一些事情。你有没有试过在记事本中将编码设置为windows-1252并导入？让我们来看看。谢谢。只需将CODEPAGE=65001添加到批量插入中，即可修复我的所有问题