Question

我正在使用pymssql将df写入mssql，但是没有用。

from sqlalchemy import create_engine
import pymssql
engine = create_engine('mssql+pymssql://sa:suzhou@localhost/test_python')

data.to_sql('phill',engine)

以下错误。怎么了？

NoSuchColumnError: "Could not locate column in row for column '0'"

TOP 5 ROWS：

    Dc_Dist Psa Dispatch_Date_Time  Dispatch_Date   Dispatch_Time   Hour    Dc_Key  Location_Block  UCR_General Text_General_Code   Police_Districts    Month   Lon Lat
3   35  D   2009-07-19 01:09:00 2009-07-19  01:09:00    1   200935061008    5500 BLOCK N 5TH ST 1500.0  Weapon Violations   20.0    2009-07 -75.130477  40.036389
4   9   R   2009-06-25 00:14:00 2009-06-25  00:14:00    0   200909030511    1800 BLOCK WYLIE ST 2600.0  All Other Offenses  8.0 2009-06 -75.166350  39.969532
5   17  1   2015-04-25 12:50:00 2015-04-25  12:50:00    12  201517017705    800 BLOCK S BROAD ST    600.0   Thefts  13.0    2015-04 -75.166412  39.940070
6   23  K   2009-02-10 14:33:00 2009-02-10  14:33:00    14  200923006310    2200 BLOCK RIDGE AVE    800.0   Other Assaults  16.0    2009-02 -75.171149  39.979586
12  22  3   2015-10-06 18:18:00 2015-10-06  18:18:00    18  201522089374    1500 BLOCK N 15TH ST    600.0   Thefts  16.0    2015-10 -75.160229  39.976134

Answer 1

猜测SQL Server不喜欢像0这样的列名，因此在将DF写入SQL Server之前必须重命名列。

所以你可以尝试下面的解决方案：

data.add_prefix('col_').to_sql('phill',engine)

Answer 2

from sqlalchemy import create_engine
import pyodbc
import pandas as pd
import numpy as np

df = pd.DataFrame(index=range(3), columns=range(3), data=np.zeros((3,3)))
engine = create_engine("mssql+pyodbc://.\SQLEXPRESS/test_python?driver=SQL server")

df.to_sql('test', con=engine)

使用本地服务器但使用pyodbc可以正常工作。您的数据框架是什么样的？

Answer 3

这是问题，因为sqlalchemy。在我的python sit-packages中，有SQLAlchemy-1.1.4-py2.7.egg-info，SQLAlchemy-1.1.5.dist-info和sqlalchemy。只需删除sqlalchemy以外的所有内容，NoSuchColumnError: "Could not locate column in row for column '0'"即可解决。

如何使用pymssql将数据帧写入mssql？

3 个答案: