pymssql 批量插入错误（python）

Question

尝试使用 pymssql 批量插入 CSV 文件，代码如下：

conn = pymssql.connect(host='server', user='user', password='secret', database='My_Dev')
cur = conn.cursor()
load = 'BULK INSERT TempStaging FROM \'/home/dross/python/scripts/var/csv/' + f + '.csv\' WITH (FIRSTROW = 1,FIELDTERMINATOR = ',',ROWTERMINATOR = \'\\n\') GO")'
cur.execute(load)

执行时得到以下错误：

Traceback (most recent call last):
  File "./uploadResults.py", line 46, in <module>
    cur.execute(sweepload)
  File "pymssql.pyx", line 447, in pymssql.Cursor.execute (pymssql.c:7092)
  File "_mssql.pyx", line 1009, in _mssql.MSSQLConnection.execute_query (_mssql.c:11585)
  File "_mssql.pyx", line 1040, in _mssql.MSSQLConnection.execute_query (_mssql.c:11459)
  File "_mssql.pyx", line 1160, in _mssql.MSSQLConnection.format_and_run_query (_mssql.c:12652)
  File "_mssql.pyx", line 203, in _mssql.ensure_bytes (_mssql.c:2733)
AttributeError: 'tuple' object has no attribute 'encode'

第 46 行是 cur.execute 行

Answer 1

请注意， .format()可能允许sql injection ，但如果您控制文件名，那么它并没有那么糟糕（不确定参数是否可以在这里工作）。

另外，在处理 SQL 时应该使用三引号字符串，你的生活会好很多。 像这样：

load = '''BULK INSERT TempStaging FROM /home/dross/python/scripts/var/csv/{}.csv WITH (FIRSTROW=1, FIELDTERMINATOR=',', ROWTERMINATOR='\n')'''.format(filename)

被三重引用，您还可以将其分解以使其更易于阅读：

load = '''
    BULK INSERT TempStaging
    FROM /home/dross/python/scripts/var/csv/{}.csv
    WITH (
        FIRSTROW=1
      , FIELDTERMINATOR=','
      , ROWTERMINATOR='\n'
    )
'''.format(filename)

Answer 2

您应该将字符串定义如下：

load = "BULK INSERT TempStaging FROM /home/dross/python/scripts/var/csv/" + f + ".csv  WITH ( FIRSTROW=1 , FIELDTERMINATOR=',' , ROWTERMINATOR='\\n')"