Question

我想使用sqlalchemy在Postgresql数据库中存储包含timedelta64类型列的熊猫数据框。阅读文档（https://docs.sqlalchemy.org/en/latest/core/type_basics.html）我希望python'timedelta'数据类型可以映射到postgresql'interval'数据类型，但是我不知道该怎么做。我尝试了以下代码：

import sqlalchemy as sa
import pandas as pd
from datetime import timedelta

engine = sa.create_engine('postgresql+psycopg2://postgres:password@floris/floris')

my_df = pd.DataFrame(data=[ timedelta(days=1), timedelta(days=2), timedelta(days=3)], index=range(0,3), columns=['delay'])

my_df.to_sql('my_table', con=engine, dtype={'delay': sa.types.Interval})

我遇到以下错误：

psycopg2.ProgrammingError: column "delay" is of type interval but expression is of type bigint
LINE 1: INSERT INTO my_table (index, delay) VALUES (0, 8640000000000...
                                                       ^
HINT:  You will need to rewrite or cast the expression.

似乎sqlalchemy并未保留timedelta数据类型，而是将其转换为bigint。该如何解决？

Answer 1

Timedelta转换为整数值。参考 https://github.com/pandas-dev/pandas/blob/master/pandas/io/sql.py#L880-L883

一种解决方法是将timedelta转换为字符串格式，然后再将其保存到db。

def a_func(val):
    return str(val)

my_df['delay'] = my_df['delay'].apply(a_func)

或

my_df['delay'] = my_df['delay'].astype(str)

如何使用sqlalchemy将具有datetime.timedelta类型数据对象的熊猫数据框存储到PostgreSQL d / b中？

1 个答案: