"TypeError: 'NoneType' object is not iterable" from pandas read_sql

Question

我正尝试在 python 中使用 pyodbc 运行 SQL 命令。其中 SQL 命令包含多个 SELECT 命令和 IF 语句。

但我收到如下错误

columns = [col_desc[0] for col_desc in cursor.description]
TypeError: 'NoneType' object is not iterable

import pyodbc
import pandas as pd
conn = pyodbc.connect("DRIVER={SQL Server};"
                "SERVER=server_name;"
                "DATABASE=master;"
                "Trusted_Connection=yes;")
cursor = conn.cursor()
script="""

If object_id ('tempdb..#Temp1')is not null
drop table #Temp1
Select distinct  a1.some_ID into #Temp1
from DOC.dbo.Document_tbl (NOLOCK)a1
from #Temp1 a1
If object_id ('tempdb..#Temp2')is not null
Drop table #Temp2
select distinct v2.some_data
into #Temp2 from tbl_name (nolock) v2 

If object_id ('tempdb..#Results')is not null
drop table #Results
select distinct a1.*,b1.####
into #Results
from #Temp1 a1
left join #Temp2 b1 on a1.## = b1.##
Select * from #Results
"""

df = pd.read_sql(script, cnxn)
writer = pd.ExcelWriter('result.xlsx')
df.to_excel(writer, sheet_name ='bar')
writer.save()

Answer 1

SQL 包含多个 SQL 语句的命令文本称为 匿名代码块 。一个匿名代码块可以 return 多个结果，其中每个结果可以是

行计数，
包含零行或多行数据的结果集，或者
一个错误。

以下示例失败...

sql = """\
SELECT 1 AS foo INTO #tmp;
SELECT * FROM #tmp;
"""
df = pd.read_sql_query(sql, cnxn)
# TypeError: 'NoneType' object is not iterable

... 因为第一个 SELECT ... INTO return 是行计数，第二个 SELECT return 是其结果集。

修复方法是使用 SET NOCOUNT ON; 启动匿名代码块，这会抑制行数并且只有 return 结果集：

sql = """\
SET NOCOUNT ON;
SELECT 1 AS foo INTO #tmp;
SELECT * FROM #tmp;
"""
df = pd.read_sql_query(sql, cnxn)
# no error

Answer 2

对于仍然出现此错误的任何其他人，我发现对于某些语句（聚合空值的 window 函数）我还需要包含 SET ANSI_WARNINGS OFF;.

"TypeError: 'NoneType' object is not iterable" from pandas read_sql

"TypeError: 'NoneType' object is not iterable" from pandas read_sql

python

sql-server

pyodbc

pandas