【转】Python例程：使用adodbapi访问MSSQL数据库

今天学习用Python访问数据库，以前用ADO习惯了，所以先找个封装了ADO的模块来试试。
adodbapi是用python封装ADO的数据库访问模块，adodbapi接口完全符合Python DB-API2.0规范，简单易用。
以下是自己学习过程中写的一个例程，使用了查询、建表、插入数据、执行存储过程等功能，基本上涵盖了日常数据库编程所需的功能。

 

#coding=utf-8

import adodbapi

class DBTestor:
    def __init__(self):
        self.conn = None
       
    def __del__(self):
        try:
            self.conn.close()
        except:
            pass
       
    def connectDB(self, connectString):
        self.conn = adodbapi.connect(connectString)
       
    def closeDB(self):
        self.conn.close()
       
    def fielddict(self, cursor):
        dict = {}
        i = 0
        for field in cursor.description:
            dict[field[0]] = i
            i += 1
        return dict
   
    def testCommand(self):
        u"测试执行SQL命令，及参数、事务"
        cursor = self.conn.cursor()
        sql = """if exists (select * from sysobjects where id = object_id(N'Demo_Table') and OBJECTPROPERTY(id, N'IsUserTable') = 1)
                    Drop Table Demo_Table;
                CREATE TABLE Demo_Table (
                   ID int IDENTITY (1, 1) NOT NULL ,
                   Name varchar(50) NOT NULL Default('')
                   PRIMARY KEY  CLUSTERED
                   (
                       [ID ]
                   )
               );"""
        cursor.execute(sql)

        sql = """INSERT INTO Demo_Table (Name) VALUES (?);"""
        cursor.execute(sql, ("jame",))
        sql = """INSERT INTO Demo_Table (Name) VALUES (?);"""
        cursor.execute(sql, ("jame2",))

        sql = """SELECT @@Identity;"""
        cursor.execute(sql)
        print "Inserted new record's ID = %s" % cursor.fetchone()[0]
       
        cursor.close()
       
        #默认对数据库进行修改后必须要提交事务，否则关闭数据库时会回滚
        self.conn.commit()

   
    def testQuery(self):
        u"测试查询功能，通过序号和字段名读取数据"
        cursor = self.conn.cursor()
        cursor.execute("SELECT * FROM authors")
        try:
            fields = self.fielddict(cursor)
            row = cursor.fetchone()
            while row != None:
                print "%s: %s %s" % (row[0], row[fields['au_fname']], row[fields['au_fname']])
                row = cursor.fetchone()                   
        finally:
            cursor.close()
           

    def testStoreProc(self):
        u"测试存储过程功能"
        cursor = self.conn.cursor()
        sql = """if exists (select * from sysobjects where id = object_id(N'insert_data_demo') and OBJECTPROPERTY(id, N'IsProcedure') = 1)
                    Drop Procedure insert_data_demo;"""
        cursor.execute(sql)
        sql = """CREATE PROCEDURE INSERT_DATA_Demo
                    @Name varchar(50),
                    @ID int output
                 AS
                    INSERT INTO Demo_Table (Name) VALUES (@Name);
                    Select @ID = @@Identity;"""
        cursor.execute(sql)
       
        (name, id) = cursor.callproc("insert_data_demo", ("tom", 0))               
        print "Inserted new record's ID = %i" % id
       
        sql = """SELECT * FROM Demo_Table;"""
        cursor.execute(sql)
        print cursor.fetchall()       
        cursor.close()
       
        self.conn.commit()
       
               
if __name__ == "__main__":
    test = DBTestor()
    test.connectDB("Provider=SQLOLEDB.1;Persist Security Info=True;Password=;User ID=sa;Initial Catalog=pubs;Data Source=.")
    try:
        test.testQuery()
        test.testCommand()
        test.testStoreProc()       
    finally:
        test.closeDB()

 

程序说明：
1. 因为源代码有中文，所以要在第一行加上 #coding=utf-8
2. adodbapi默认在Connection的__init__里面开启事务BeginTrans()，然后在close方法里面回滚事务RollbackTrans()。因此对数据库进行修改之后要调用Connection.commit()提交事务才会保存数据。这一点与ADO的默认行为不一致，需要注意。
3. cursor.fetch*()方法返回的是tuple，只能通过序号读取字段值，如果要通过字段名访问数据，需要将字段名映射为序号。测试代码里面的fielddict方法利用cursor.description属性实现了此映射过程。

总结：
感觉Python DB-API2.0接口使用是很简单和方便的，只不过接口功能稍弱了点，比如没有MoveFirst等功能，不能多次遍历单个结果记录集，而且不直接支持使用字段名访问数据，需要做转换。

参考资料：
1. PEP-0249, Python Database API Specification V 2.0
2.
3.

本文来自CSDN博客，转载请标明出处：http://blog.csdn.net/ildf/archive/2009/01/24/3851427.aspx