使用 python panda 模块附加到 sql 表

Question

I am attempting to append to a SQL table using to_sql command in python. Everything works fine but if I run my command more tahn once it will obviously append the same data. I know this could be solved with if_exists='replace' but when i do that i get the following error.

ProgrammingError: (pyodbc.ProgrammingError) ('42000', "[42000] [Microsoft][ODBC SQL Server Driver][SQL Server]Column 'AccountName' in table 'table' is of a type that is invalid for use as a key column in an index. (1919) (SQLExecDirectW)") [SQL: u'CREATE INDEX [ix_table_AccountName] ON table ([AccountName])']

I need to figure out how to either,only append data that is not already in the table, delete the data I am about to append if it is already in the table or how to fix my index problem so i can use if_exists='replace'

engine = create_engine('mssql+pyodbc://UN:PW@DB')
df.to_sql("table", engine,if_exists='replace')

Answer 1

Your column 'AccountName' is probably too long. For MSSQL the index can only be 900 bytes, so if you use utf-8 this is 450 chars.

Try setting the dtype for your column in the to_sql() statement with a shorter length if that column is used as index.

from sqlalchemy.types import String
[...]
df.to_sql(
[...]
    dtype={
       [...]
       'AccountName': String(255),