pip3 install pydbclib

from pydbclib import connect
# 使用with上下文，可以自动提交，自动关闭连接
with connect("sqlite:///:memory:") as db:
    db.execute(‘create table foo(a integer, b varchar(20))‘)
    # 统一使用’:[name]‘形式的SQL的占位符
    db.execute("insert into foo(a,b) values(:a,:b)", [{"a": 1, "b": "one"}]*4)
    print(db.read("select * from foo").get_one())
    print(db.read("select * from foo").get_all())
    print(db.read("select * from foo").to_df())
    db.get_table("foo").insert({"a": 2, "b": "two"})
    print(db.get_table("foo").find_one({"a": 2}))
    print(db.get_table("foo").find().get_all())
    print(db.get_table("foo").find().to_df())

# connect函数有个driver参数决定你是通过哪个数据库驱动包去连接的
# driver参数默认值是sqlalchemy，即通过sqlalchemy驱动包连接数据库
>>> db = pydbclib.connect("sqlite:///:memory:")
>>> db = pydbclib.connect(":memory:", driver=‘sqlite3‘)
# 也可以传入驱动包连接对象
>>> import sqlite3
>>> db = pydbclib.connect(driver=sqlite3.connect(":memory:"))
>>> from sqlalchemy import create_engine
>>> db = pydbclib.connect(driver=create_engine("sqlite:///:memory:"))

>>> record = {"a": 1, "b": "one"}
>>> db.execute(‘create table foo(a integer, b varchar(20))‘)
# 插入单条记录，结果返回影响行数
>>> db.execute("insert into foo(a,b) values(:a,:b)", record)
1
# 插入多条记录
>>> db.execute("insert into foo(a,b) values(:a,:b)", [record, record])
2

# 查询结果只返回一条记录
>>> db.read_one("select * from foo")
{‘a‘: 1, ‘b‘: ‘one‘}
#read返回迭代器类型，用get方法获取前几条记录，使用map对每条记录进行数据清洗
>>> db.read("select * from foo").map(lambda x: {f"foo.{k}": v for k,v in x.items()}).get(2)
# as_dict=False返回原生元祖记录
>>> db.read("select * from foo", as_dict=False).get(2)
[(1, ‘one‘), (1, ‘one‘)]
# 也可以直接for遍历
>>> for r in db.read("select * from foo"):
...     print(r)
... 
{‘a‘: 1, ‘b‘: ‘one‘}
{‘a‘: 1, ‘b‘: ‘one‘}
{‘a‘: 1, ‘b‘: ‘one‘}
# 转换成pandas dataframe对象, 前提已经安装了pandas
>>> db.read("select * from foo").to_df()
   a    b
0  1  one
1  1  one
2  1  one

>>> db.rollback()
>>> db.commit()
>>> db.close()

# 插入单条和插入多条，输入参数字典的键值必须和表中字段同名
>>> db.get_table("foo").insert({"a": 1, "b": "one"})
1
>>> db.get_table("foo").insert([{"a": 1, "b": "one"}]*10)
10

# 查询字段a=1第一条记录
>>> db.get_table("foo").find_one({"a": 1})
{‘a‘: 1, ‘b‘: ‘one‘}
# 也可以直接写成sql条件表达式，其他接口的条件参数类似都可以是表达式
>>> db.get_table("foo").find_one("a=1")
{‘a‘: 1, ‘b‘: ‘one‘}
# 查询字段a=1所有记录，find返回迭代器对象同上面read方法
>>> db.get_table("foo").find({"a": 1}).get_all()
[{‘a‘: 1, ‘b‘: ‘one‘},...{‘a‘: 1, ‘b‘: ‘one‘}]

# 将a=1那条记录的b字段值更新为"first"
>>> db.get_table("foo").update({"a": 1}, {"b": "first"})
11
>>> db.get_table("foo").find({"a": 1}).get_one()
{‘a‘: 1, ‘b‘: ‘first‘}

# 将a=1那条记录删除
>>> db.get_table("foo").delete({"a": 1})
11
>>> db.get_table("foo").find({"a": 1}).get_all()
[]

# 使用普通数据库驱动连接，driver参数指定驱动包名称
# 例如pymysql包driver=‘pymysql‘,connect函数其余的参数和driver参数指定的包的创建连接参数一致
# 连接mysql
db = pydbclib.connect(user="user", password="password", database="test", driver="pymysql")
# 连接oracle
db = pydbclib.connect(‘user/password@local:1521/xe‘, driver="cx_Oracle")
# 通过odbc方式连接
db = pydbclib.connect(‘DSN=mysqldb;UID=user;PWD=password‘, driver="pyodbc")  
# 通过已有驱动连接方式连接
import pymysql
con = pymysql.connect(user="user", password="password", database="test")
db = pydbclib.connect(driver=con)

# 使用Sqlalchemy包来连接数据库，drvier参数默认为‘sqlalchemy‘
# 连接oracle
db = pydbclib.connect("oracle://user:password@local:1521/xe")
# 连接mysql
db = pydbclib.connect("mysql+pyodbc://:@mysqldb")
# 通过已有engine连接
from sqlalchemy import create_engine
engine = create_engine("mysql+pymysql://user:password@localhost:3306/test")
db = pydbclib.connect(driver=engine)