1) Installation

Open a terminal and install pymy package

pip install pymy

2) Use

1. Be sure that you have set environment variables with Redshift credentials like this:

export MYSQL_{INSTANCE}_DATABASE=""
export MYSQL_{INSTANCE}_USERNAME=""
export MYSQL_{INSTANCE}_HOST=""
export MYSQL_{INSTANCE}_PORT=""
export MYSQL_{INSTANCE}_PASSWORD=""

2. Be also sure that your IP address is authorized for the mysql cluster/instance.

3. Prepare your data like that:

data = {
        "table_name"    : 'name_of_the_redshift_schema' + '.' + 'name_of_the_redshift_table'
        "columns_name"  : [first_column_name,second_column_name,...,last_column_name],
        "rows"      : [[first_raw_value,second_raw_value,...,last_raw_value],...]
    }

4. Send your data (use the same {INSTANCE} parameter as environment variables):

import pymy
pymy.send_to_redshift(instance, data, replace=True, batch_size=1000, types=None, primary_key=(), create_boolean=False)

• replace (default=True) argument allows you to replace or append data in the table

• batch_size (default=1000) argument also exists to send data in batchs

• types, primary_key and create_boolean are explained below

3) First Example

You have a table called dog in you animal scheme. This table has two columns : ‘name’ and ‘size’. You want to add two dogs (= two rows) : Pif which is big and Milou which is small. data will be like that:

import pymy
data = {
        "table_name"    : 'animal.dog'
        "columns_name"  : ['name','size'],
        "rows"      : [['Pif','big'], ['Milou','small']]
    }
pymy.send_to_redshift({INSTANCE},data)

4) Function create_table

pymy has a create_table function with this signature:

import pymy
pymy.create_table({INSTANCE}, data, primary_key=(), types=None)

This function is automatically called in the send_to_redshift function if the table is not created. You can also call it with the “create_boolean” parameter of the send_to_reshift function or by setting “primary_key” or “types” parameters.

• primary_key : if you have 3 columns (ie: columns_name=[a,b,c]) and you want to set b as primary key, set primary_key=(b)

• types: create_table function guesses types of each column. But you can set a “types” argument. It is a dictionary like {‘b’: ‘VARCHAR(12)’} or {‘b’: ‘INTEGER NOT NULL’} to set types of b column.