ClickSQL 0.1.8.9.2

SQL programming

ClickSQL: ClickHouse client for Humans

Package information:

ClickSQL is a smart client for ClickHouse database, which may help users to use ClickHouse more easier and pythonic. More information for ClickHouse can be found at here

Installation

pip install ClickSQL

Usage

Initial connection

to setup a database connection and send a heartbeat-check signal

from ClickSQL import BaseSingleFactorTableNode

conn_str = "clickhouse://default:test121231@99.99.9.9:8123/system"
Node = BaseSingleFactorTableNode(conn_str)

>>> connection test:  Ok.

Query

execute a SQL Query

from ClickSQL import BaseSingleFactorTableNode

conn_str = "clickhouse://default:test121231@99.99.9.9:8123/system"
Node = BaseSingleFactorTableNode(conn_str)

Node('show tables from system limit 1')

>>> connection test:  Ok.
>>>                             name
>>> 0  aggregate_function_combinators

execute a Query without SQL

from ClickSQL import BaseSingleFactorTableNode

factor = BaseSingleFactorTableNode(
        'clickhouse://default:default@127.0.0.1:8123/sample.sample',
        cols=['cust_no', 'product_id', 'money'],
        order_by_cols=['money asc'],
        money='money >= 100000'
    )


factor['money'].head(10)

>>> connection test:  Ok.
>>>        money
>>> 0  1000000.0
>>> 1  1000000.0
>>> 2  1000000.0
>>> 3  1000000.0
>>> 4  1000000.0
>>> 5  1000000.0
>>> 6  1000000.0
>>> 7  1000000.0
>>> 8  1000000.0
>>> 9  1000000.0

Insert data

insert data into database by various ways

Insert data via DataFrame

from ClickSQL import BaseSingleFactorTableNode as factortable
import numpy as np
import pandas as pd
factor = factortable( 'clickhouse://default:default@127.0.0.1:8123/sample.sample'  )
db = 'sample'
table = 'sample'
df  = pd.DataFrame(np.random.random(size=(10000,3)),columns=['cust_no', 'product_id', 'money'])
factor.insert_df(df, db, table, chunksize=100000)

Insert data via SQL(Inner)

from ClickSQL import BaseSingleFactorTableNode as factortable

factor = factortable( 'clickhouse://default:default@127.0.0.1:8123/sample.sample'  )

factor("insert into sample.sample select * from other_db.other_table")

Create table

Create table by SQL

from ClickSQL import BaseSingleFactorTableNode

conn_str = "clickhouse://default:test121231@99.99.9.9:8123/system"
Node = BaseSingleFactorTableNode(conn_str)

Node('create table test.test2 (v1 String, v2 Int64, v3 Float64,v4 DataTime) Engine=MergeTree() order by v4')

Create table by DataFrame

from ClickSQL import BaseSingleFactorTableNode
import numpy as np
import pandas as pd

conn_str = "clickhouse://default:test121231@99.99.9.9:8123/system"
Node = BaseSingleFactorTableNode(conn_str)
db = 'test'
table = 'test2'


df_or_sql_or_dict  = pd.DataFrame(np.random.random(size=(10000,2)),columns=['v1', 'v3'])
df_or_sql_or_dict['v2'] =1
df_or_sql_or_dict['v4'] =pd.to_datetime('2020-01-01 00:00:00')

Node.create( db,  table,  df_or_sql_or_dict,    key_cols=['v4'],)

Contribution

Welcome to improve this package or submit an issue or any others

Author

sn0wfree

Plan

Available functions or properties

1. get data from clickhouse
2. insert data into clickhouse
3. create
4. alter
5. execute standard SQL and transform into DataFrame(Auto)
6. able to execute select query
7. able to execute insert query
8. no require clickhouse-client
9. auto create table sql
10. can execute explain query
11. Insert Data via DataFrame
12. alter function & drop function

In Process

2. create a pandas_liked executable function, which can compatible with pandas
3. distributed query（query+insert）

schedule

1. ORM
2. can execute user role query
3. create analysis component
4. auto report system
5. table register system
6. data manager system
7. meta data manager