安装Duckdb
DuckDB 是一个高性能的分析数据库系统,可以不用启动启动服务端,临时在内存中起个服务,支持sql语法
pip install duckdb
与pandas结合使用
import duckdb
import pandas as pd
df = pd.DataFrame(
[
[1, "吃饭"],
[1, "蔡英文"],
[1, "睡觉"],
[1, "打豆豆"],
[2, "吃饭"],
[2, "打豆豆"],
[2, "山东"],
[3, "吃饭"],
[3, "河南"],
[3, "北京"],
[3, "上海"]
], columns = ['key', 'act'])
with duckdb.connect() as con:
con.register("df_test", df)
results = con.execute("SELECT key, count(act) as count from df_test group by key").df().to_dict("records")
print(results)
从示例看duckdb与pandas结合使用非常简单,支持sql语法,这让pandas的使用更加灵活。