ThinkSQL 类似 ThinkPHP 的数据库引擎

Drivers

• MySQL
• 达梦(DM8)

Install

• use mysql

pip install think_sql[mysql]

• use 达梦(DM8)

pip install think_sql[dm]

How to use

1. simple demo

Database: test Table: user

• example dict params

from think_sql import DB

config = {
    'type': 'mysql',
    'host': '127.0.0.1',
    'port': 3306,
    'user': 'root',
    'password': 'root',
    'database': 'test',
}

with DB(config) as db:
    data = db.table('user').where('id',1).find()
    print(data)

• example dsn str

from think_sql import DB

with DB("mysql://root:'root'@127.0.0.1:3306/test") as db:
    data = db.table('user').where('id',1).find()
    print(data)

• example DBConfig

from think_sql import DB
from think_sql.tool.util import DBConfig
config = DBConfig(
  type='mysql',
  host='127.0.0.1',
  port=3306,
  user='root',
  password='root',
  database='test',
)
with DB(config) as db:
    data = db.table('user').where('id',1).find()
    print(data)

result

{
  "id": 1,
  "username": "hbh112233abc",
  "age": "36",
  "address": "FUJIAN.XIAMEN"
}

• use db function

from think_sql import db

config = {
    'type': 'mysql',
    'host': '127.0.0.1',
    'port': 3306,
    'user': 'root',
    'password': 'root',
    'database': 'test',
}

conn = db(config)
data = conn.table('user').where('id',1).find()
print(data)
conn.close()

2. Introduction

DB

• init(config:Union[str,dict,DBConfig],params={})

  init database, return DB instance

  • config:Union[str,dict,DBConfig]
    • str: type://user:'password'@host:port/database
    • dict: {'type':'mysql','host':'127.0.0.1','port':3306,'user':'root','password':'root','database':'test'}
    • DBConfig: DBConfig(type='mysql',host='127.0.0.1',port=3306,user='root',password='root',database='test')
  • params:dict pymysql connect other params

• connect() connect database use init params

• table(table_name):Table return class Table <think_sql.tool.interface.TableInterface>

• check_connected():bool check connected, try reconnect database

• query(sql,params=()) query sql return cursor.fetchall List[dict]

• execute(sql,params=()) execute sql write operate(ex:insert,update,delete,...)

Table

• init(connector: Connection,cursor: Cursor,table_name: str,debug: bool = True)

• init() initialize query condition

• debug(flag=True) set debug flag

• set_cache_storage(storage: CacheStorage) set cache storage ex: Redis

• cache(key: str = None, expire: int = 3600) use cache at query

• cursor(sql: str, params: list = []) -> Cursor return cursor object

• get_last_sql() -> str return last sql string

• get_lastid() -> str return last row id

• get_rowcount() -> int return affect rows count

• fetch_sql(flag: bool = True) set fetch sql flag,if flag = True then query and execute will only return sql

• build_sql(operation: str, params: list = []) -> str return build sql

• query(sql: str, params: list = []) -> list execute read operation sql and return cursor.fetchall() when fetch_sql=True then return sql and not execute the sql

• execute(sql: str, params: list = []) -> int execute write operation sql and return affect rows count when fetch_sql=True then return sql and not execute the sql

• where(field: Union[str, list, tuple], symbol: str = '', value: Any = None) set query conditions, support multipe use

  where(field,value)

  where field = value
  

  where(field,symbol,value)

  where field symbol value
  

  where(
      [
          [field1,symbol1,value1],
          [field2,symbol2,value2]
      ]
  )
  

  where field1 symbol1 value1 and field2 symbol2 value2
  

  where(field1,symbol1,value1).where(field2,symbol2,value2)

  where field1 symbol1 value1 and field2 symbol2 value2
  

  • symbol

  symbol	another	demo
  =	eq,=	where('id','=',1)
  <>	neq, !=, <>	where('id','<>',1)
  >	gt,>	where('id','>',1)
  >=	egt,>=	where('id','>=',1)
  <	lt, <	where('id','<',1)
  <=	elt,<=	where('id','<=',1)
  in	in,not in	where('id','in',[1,2,3])
  between	between,not between	where('id','between',[1,5]) where('id','between','1,5') where('id','between','1 and 5')
  like	like, not like	where('name','like','%hbh%')
  null	is null,null	where('remark','is null')
  not null	is not null,not null	where('remark','is not null')
  exists	exists, not exists	where('remark','exists')
  exp	exp	where('id','exp','in (1,2,3)')

• where_or(field: Union[str, list], symbol: str = '', value: Any = None)

  where('id',1).where_or('id',5)

  where id = 1 or id = 5
  

• limit(start: int, step: int = None) LIMIT start,step

• page(index: int = 1, size: int = 20) LIMIT index*size-1,size

• order(field: str, sort: str = 'asc') ORDER BY field sort

• group(field:str) GROUP BY field

• distinct(field:str) SELECT DISTINCT field

• field(fields: Any, exclude: bool = False) SELECT fields if exclude=True then select the fields of table (exlude:fields)

• select(build_sql: bool = False) -> list return select query result if build_sql=True then return sql

• find() return select ... limit 1

• value(field: str) return the field of first row

• column(field: str,key: str = '')

  column('name')

  return ['hbh','mondy']

  column('name,score')

  return [{'hbh':80},{'mondy':88}]

  column('score','name')

  return {'hbh':80, 'mondy':88}

  column('id,score','name')

  return { 'hbh':{'id':1,'score':80}, 'mondy':{'id':2,'score':88} }

• alias(short_name: str = '') set alias table_name

• join(table_name: str, as_name: str = '', on: str = '', join: str = 'inner', and_str: str = '')

  • table_name join table_name
  • as_name alias short_table_name for table_name
  • on join condition
  • join join type in 'INNER', 'LEFT', 'RIGHT', 'FULL OUTER'
  • and_str and condition demo

  db.table('table1').alias('a').join(
    'table2','b','a.id=b.a_id','left'
  ).join(
    'table3','c','c.a_id=a.id'
  ).field(
    'a.id,a.name,b.id as b_id,b.score,c.id as c_id,c.remark'
  ).where(
    'a.id',1
  ).find()
  

  sql

  SELECT
      a.id,
      a.name,
      b.id AS b_id,
      b.score,
      c.id AS c_id,
      c.remark
  FROM
      table1 AS a
      LEFT JOIN table2 AS b ON a.id = b.a_id
      INNER JOIN table3 AS c ON c.a_id = a.id
  WHERE
      a.id = 1
      LIMIT 1
  

• union(sql1: str, sql2: str, union_all: bool = False) union sql1 and sql2

  • union_all if union_all is True then UNION ALL

  demo

  sql1 = db.table('table1').field('name,score').where('status',1).select(build_sql=True)
  sql2 = db.table('table2').field('name,score').where('status',1).select(build_sql=True)
  
  result = db.table().union(sql1,sql2).where('score','>',60).select()
  

  sql

  SELECT
  *
  FROM
      ( SELECT `name`, `score` FROM table1 WHERE `status` = 1 )
      UNION
      ( SELECT `name`, `score` FROM table2 WHERE `status` = 1 )
  WHERE
      score > 60
  

• insert(data: Union[dict, List[dict]], replace: bool = False) -> int insert data to database

  • data dict: insert one record; list: insert multiple records
  • replace bool if replace is True then use REPLACE INTO

  demo insert one record

  db.table('table1').insert({'name':'test','score':100})
  

  INSERT INTO table1 (`name`, `score`) VALUES ('test', 100)
  

  insert multiple records

  db.table('table1').insert([{'name':'test','score':100},{'name':'test2','score':101}])
  

  INSERT INTO table1 (`name`, `score`) VALUES ('test', 100), ('test2', 101)
  

  replace mode

  db.table('table1').insert({'id':1,'name':'test','score':100},replace=True)
  

  REPLACE INTO table1 (`id`, `name`, `score`) VALUES (1,'test', 100)
  

• update(data: dict, all_record: bool = False) -> int update data

  • data dict you want update data
  • all_record bool if all_record is False then you must set update condition; if you want to update all records then you need set all_record = True

• delete(all_record: bool = False) -> int delete record

  • all_record bool if all_record is False then you must set delete condition; if you want to delete all records then you need set all_record = True

• inc(field: str, step: Union[str, int, float] = 1) -> int

  increase field +step

• dec(field: str, step: int = 1) -> int

  decrease field -step

• max(field: str) -> Union[int, float]

  get the max value of field

• sum(field: str) -> Union[int, float, Decimal]

  get the sum value of field

• avg(field: str) -> Union[int, float, Decimal]

  get the avg value of field

• count(field: str = '*') -> int

  get the count of records

• copy_to(new_table: str = None, create_blank_table: bool = False) -> int

  copy data to new_table

  • new_table if new_table is None then new_table will auto set like {table_name}_copy
  • create_blank_table bool if create_blank_table is True then only create a blank table like current table.

  demo

  db.table('user').field(
    'name,score'
  ).where(
    'score','>',60
  ).copy_to('good_boy')
  

  sql

  SELECT
    `name`,
      `score`
  INTO `good_boy`
  FROM
  `user`
  WHERE
      score > 60
  

• insert_to(new_table: str, fields: Union[str, list, tuple] = None) -> int

  INSERT INTO {new_table} SELECT {select_fields} FROM {table} {join} WHERE {where}{group}{order}{limit}
  

• exists(self) -> bool

  check record exists with some query conditions, it use SELECT 1 FROM {table} {join} WHERE {where} LIMIT 1

• batch_update(data:List[dict],key:str) -> int

  batch update multiple records

  demo

  data = [
      {'id':1,'score':66},
      {'id':2,'score':59},
      {'id':3,'score':86},
      {'id':4,'score':90},
  ]
  db.table('user').batch(data,key='id')
  

  sql

  update `user` set score = 66 where id = 1;
  update `user` set score = 59 where id = 2;
  update `user` set score = 86 where id = 3;
  update `user` set score = 90 where id = 4;
  

support transaction

from think_sql import DB
db_dsn = "root:'password'@127.0.0.1:3306/database"
with DB(db_dsn) as db:
    # result: insert two records into database
    with db.start_trans():
        db.table('user').insert({'name':'think_sql1','score':98})
        db.table('user').insert({'name':'think_sql2','score':99})
    # result: nothing inserted
    with db.start_trans():
        db.table('user').insert({'name':'think_sql1','score':98})
        db.table('user').insert({'name':'think_sql2','score':99})
        raise Exception('error')

    # The above operation does not affect subsequent operations.
    db.table('user').insert({'name':'think_sql3','score':100})

sql_helper for mysql

Ref:hcymysql/sql_helper

from think_sql import DB
from think_sql.mysql.sql_helper import help

db_dsn = "root:'password'@127.0.0.1:3306/database"
with DB(db_dsn) as db:
    sql = "slow query sql"
    help(db, sql)

result

1) 输入的SQL语句是：
----------------------------------------------------------------------------------------------------
SELECT *
FROM hy_cabrecs
WHERE finished_count > 0
----------------------------------------------------------------------------------------------------
2) EXPLAIN执行计划:
+------+---------------+------------+--------------+--------+-----------------+-------+-----------+-------+--------+------------+-------------+
| id   | select_type   | table      | partitions   | type   | possible_keys   | key   | key_len   | ref   | rows   | filtered   | Extra       |
+======+===============+============+==============+========+=================+=======+===========+=======+========+============+=============+
| 1    | SIMPLE        | hy_cabrecs | None         | ALL    | None            | None  | None      | None  | 14422  | 33.33      | Using where |
+------+---------------+------------+--------------+--------+-----------------+-------+-----------+-------+--------+------------+-------------+
3) 索引优化建议：
----------------------------------------------------------------------------------------------------
取出表 【hy_cabrecs】 where条件字段 【finished_count】 100000 条记录，重复的数据有：【16093】 条，没有必要为该字段创建索引。
 【hy_cabrecs】 表，无需添加任何索引。

parse for mysql

• parse alter sql

  from think_sql.mysql import parse
  sql = """
    alter     table
    slow_log_test
    add
    iii int not null default 0  comment   'a';
  """
  
  print(parse.alter_sql(sql))
  

  result:

  {
    "table": "slow_log_test",
    "field": "iii",
    "field_type": "int",
    "is_null": "NOT NULL",
    "default": "0",
    "comment": "'a'",
    "after": ""
  }
  

diff for mysql

比较两个mysql实例的差异并生成update.sql

python -m think_sql.mysql.diff --type mysql --src "root:'root'@127.0.0.1:3306/db1" --dst "root:'xmhymake'@192.168.102.154:3306/db2"

Development

poetry 包管理器

官网

Python 包管理之 poetry 的使用

Python 包管理之 poetry 基本使用

# 配置虚拟环境在项目目录下
poetry config virtualenvs.path true
# 安装依赖
poetry install
# 进入虚拟环境
poetry shell

poetry command

unit test

pytest --cov --cov-report=html

publish

poetry build
poetry config pypi-token.pypi "your pypi.org api token"
poetry publish -n