Use Sqlalchemy module to obtain the table name and field name of the existing table in the database

Recently used Python in ORM modular , for example SQLalchemy To read tables and related data in the database , At first contact , Most of the time, they imitate other bloggers or official cases on the Internet , Fine tune the code for the database you need to operate , Then I have some experience in the process of use , Close up this article only for experience sharing .

/ Module and database

• windows 10

• python 3.7

• sqlalchemy 1.3.13

• mysql-connector 8.0.28

• mysql 8.0.18

/ brief introduction

By comparison operation mysql Under the sql Grammar and orm The similarities and differences of grammar learning in reading data tables .

/ mysql grammar

Connect to database ：

Use the command line to enter the database , If not mysql.exe Set in the system environment , Please switch the command line path to mysql.exe Location path ,Ctrl+R Then type in cmd Go to the command line .

mysql -u root -p
Enter password: password

Display all databases ：

show databases;

Using a database ：

use database;  -- database:  Enter your own database name 

Query the data of the data table ：

select * from table; -- table:  Target data table 

The above is the use of sql Syntax uses the database to query the data of the data table under the current table .

/ ORM(SQLalchemy) The way

Use sqlalchemy when , If there is no corresponding table in the current library, a new table model class will be created , Take this as the query condition , Of course, it has been explained in this article that there are tables , Of course, you can complete the operation of obtaining data tables without establishing model classes .

I found it useful this time pymysql As a connector, there are always warnings , Switch to mysqlconnector Just fine .

pip intall mysql-connector-python

Create table model classes

from sqlalchemy import create_engine, MetaData, Column
from sqlalchemy.dialects.mysql import INTEGER, DOUBLE, BIGINT, VARCHAR, CHAR, TEXT, DATETIME
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import sessionmaker

#  Create database engine ,ps: There is no connection to the specific database 
engine = create_engine("mysql+mysqlconnector://root:[email protected]:3306", pool_recycle=7200)
#  Create base class ,  Use MetaData Module building library model metadata class 
Base = declarative_base(bind=engine, metadata=MetaData(schema='dash_base'))

#  Use sessionmaker Keep the database session connected 
session = sessionmaker(bind=engine)()

class Tips(Base):
    """ Inherited base class , Design the specific column names or formats in the table """
    __tablename__ = 'tips'
    idx = Column('idx', BIGINT(20), primary_key=True)
    total_bill = Column('total_bill', DOUBLE())
    tip = Column('tip', DOUBLE())
    sex = Column('sex', TEXT())
    smoker = Column('smoker', TEXT())
    day = Column('day', TEXT())
    time = Column('time', TEXT())
    size = Column('size', BIGINT(20))

Query all the data

[tips.__dict__ for tips in session.query(Tips).all()]

Query partial column data

session.query(Tips.total_bill, Tips.tip).all()

Sometimes it will combine pandas Use it together .

import pandas as pd

# table name ,con： The link of parameters should be specific to the database name 
pd.read_sql(Tips.__tablename__, con="mysql+mysqlconnector://root:[email protected]:3306/dash_base")

（ Manual watermark ： original CSDN The fate of the sleepers ,https://blog.csdn.net/weixin_46281427?spm=1011.2124.3001.5343, official account A11Dot send )
There is no difficulty in understanding the above code , The main reason is that you need to manually type in the table model class every time , It will be troublesome to write the existing table like this again , If the number of table fields increases , The number of tables to query increases , Will want to find a better solution , Of course there are , Please read on .

Use insepctor Module and MetaData class

insepctor There are two ways to create ：

• Mode one

from sqlalchemy.engine import reflection

engine = create_engine("mysql+mysqlconnector://root:[email protected]:3306", pool_recycle=7200)
insp = reflection.Inspector.from_engine(engine)

• Mode two

from sqlalchemy import create_engine, inspect, MetaData, Table
from sqlalchemy.orm import sessionmaker

engine = create_engine("mysql+mysqlconnector://root:[email protected]:3306", pool_recycle=7200)
insp = inspect(engine)

Get database name

insp.get_schema_names()

Get the table name under the current database

insp.get_table_names(schema='dash_base')  # schema:  Database name 

Get the column name of the current data table

inspect_resp.get_columns('tips', schema='dash_base')  #  Table name , Library name 

name Value is column name ,type Value is the type of column . However, this method of obtaining data is not helpful for querying the data in the table , The reason is that it cannot be used str Type as session.query() The object of .

Next use MetaData Class connection database mapping the table structure in the delivery .

Get the database name in the previous step , Select a database name and fill it in schema Parameters in .

metadata = MetaData(bind=engine)
# reflect  mapping dash_base Table structure under the Library 
metadata.reflect(schema='dash_base')

After passing this code , You can use the following tables Method to get the table name and table structure ：

Table name ：

Table structure ：

You can also use Table Class to create a new table structure ：

from sqlalchemy import Table

Table('tips', metadata, schema='dash_base')

I recommend using a dictionary in the form of metadata.tables Analyze it , Convenient retrieval , You don't have to use it anymore Table() class .

table_dicts = {
    i.name: i for i in metadata.tables.values()}

And then get the Field name ：

table_dicts['tips'].c.keys()

# ['idx', 'total_bill', 'tip', 'sex', 'smoker', 'day', 'time', 'size']

The field name obtained in this way uses session.query A query is a list of data , Instead of iterators .

Say an extra sentence , have access to Base.metadata Instead of MetaData, Same effect .

summary

Every time I use a module, I will feel a little more , Only by recording in time can we better understand and use this module , At first, every time I want to query through this module, I have to create a table model class , Also on the other hand, spit out that the design of this module is unreasonable , If you don't know the specific design of the table structure, how to query , Through the accumulation of experience , Slowly feel orm The model design is so robust , So convenient .

The world is full of flowers and grass , Why not raise a glass and cheer for the endless beauty ！

---

Made on February 14th, 2002