SQL IN operator using pyodbc and SQL Server

问题

I'm using pyodbc to query to an SQL Server database

import datetime
import pyodbc    
conn = pyodbc.connect("Driver={SQL Server};Server='dbserver',Database='db',
                       TrustedConnection=Yes")
cursor = conn.cursor()
ratings = ("PG-13", "PG", "G")
st_dt = datetime(2010, 1, 1)
end_dt = datetime(2010, 12, 31)
cursor.execute("""Select title, director, producer From movies 
                Where rating In ? And release_dt Between ? And ?""", 
                ratings, str(st_dt), str(end_dt))

but am receiving the error below. Does the tuple parameter need to be handled in a different way? Is there a better way to structure this query?

('42000', "[42000] [Microsoft][ODBC SQL Server Driver][SQL Server]Line 9: 
  Incorrect syntax near '@P1'. (170) (SQLExecDirectW); 
  [42000] [Microsoft][ODBC SQL Server Driver][SQL Server]
  Statement(s) could not be prepared. (8180)")

UPDATE:

I was able to get this query to work using the string formatting operator, which isn't ideal as it introduces security concerns.

import datetime
import pyodbc    
conn = pyodbc.connect("Driver={SQL Server};Server='dbserver',Database='db',
                       TrustedConnection=Yes")
cursor = conn.cursor()
ratings = ("PG-13", "PG", "G")
st_dt = datetime(2010, 1, 1)
end_dt = datetime(2010, 12, 31)
cursor.execute("""Select title, director, producer From movies 
                Where rating In %s And release_dt Between '%s' And '%s'""" % 
                (ratings, st_dt, end_dt))

回答1:

You cannot parameterize multiple values in an IN () clause using a single string parameter. The only way to accomplish that is:

String substitution (as you did).
Build a parameterized query in the form IN (?, ?, . . ., ?) and then pass in a separate parameter for each place holder. I'm not an expert at Python to ODBC but I imagine that this is particularly easy to do in a language like Python. This is safer because you get the full value of parameterization.

回答2:

To expand on Larry's second option - dynamically creating a parameterized string, I used the following successfully:

placeholders = ",".join("?" * len(code_list))
sql = "delete from dbo.Results where RESULT_ID = ? AND CODE IN (%s)" % placeholders
params = [result_id]
params.extend(code_list)
cursor.execute(sql, params)

Gives the following SQL with the appropriate parameters:

delete from dbo.Results where RESULT_ID = ? AND CODE IN (?,?,?)

回答3:

The problem is your tuple. The ODBC connection is expecting a string to construct the query and you are sending a python tuple. And remember that you have to get the string quoting correct. I'm assuming that the number of ratings you will be looking for varies. There is probably a better way, but my pyodbc tends to be simple and straightforward.

Try the following:

import datetime
import pyodbc    
conn = pyodbc.connect("Driver={SQL Server};Server='dbserver',Database='db',
                       TrustedConnection=Yes")

def List2SQLList(items):
    sqllist = "%s" % "\",\"".join(items)
    return sqllist


cursor = conn.cursor()
ratings = ("PG-13", "PG", "G")
st_dt = datetime(2010, 1, 1)
end_dt = datetime(2010, 12, 31)
cursor.execute("""Select title, director, producer From movies 
                Where rating In (?) And release_dt Between ? And ?""", 
                List2SQLList(ratings), str(st_dt), str(end_dt))

回答4:

To expand on Larry and geographika's answers:

ratings = ('PG-13', 'PG', 'G')
st_dt = datetime(2010, 1, 1)
end_dt = datetime(2010, 12, 31)

placeholders = ', '.join('?' * len(ratings))
vars = (*ratings, st_dt, end_dt)
query = '''
    select title, director, producer
    from movies
    where rating in (%s)
       and release_dt between ? and ?
''' % placeholders

cursor.execute(query, vars)

With the placeholder, this will return a query of:

    select title, director, producer
    from movies
    where rating in (?, ?, ?)
       and release_dt between ? and ?

If you pass in ratings, it'll attempt to fit all of its items into one ?. However, if we pass in *ratings, and each item in ratings will take its place in the in() clause. Thus, we pass the tuple (*ratings, st_dt, end_dt) to cursor.execute().

来源：https://stackoverflow.com/questions/4819356/sql-in-operator-using-pyodbc-and-sql-server

标签

python

sql

pyodbc