Snowflake introduction

Content for the Snowflake introduction course

Preparation before the course

Create Snowflake trial
Take Snowflake for a spin
Install Python
Install Docker Desktop
Install git

Create Snowflake trial

Go to signup.snowflake.com and create a trial

Choose

Enterprise (the higher the tier, the more features and more expensive compute)
AWS (newer Snowflake features are rolled out on AWS before other platforms)
Location (just what is closest to you)

Activate your account via your e-mail

Write down:

User name
Password
Link to your account (the part before ".snowflakecomputing.com" is the important part)
- Example: xy97133.eu-central-1.snowflakecomputing.com

Take Snowflake for a spin

Click on Worksheets
Click on + Worksheet

Run the following code:

select top 100 
    *
from
    snowflake_sample_data.tpch_sf10.orders;

To run code, hit the run button in the top left:

If you see this, congratulations you are a now a Snowflake user ❄️💪

50,400,240,948

To feel the power of Snowflake.

use schema snowflake_sample_data.tpcds_sf10tcl; --50b
use schema snowflake_sample_data.tpcds_sf10tcl; --500b

alter session set use_cached_result = true;

-- Count
with row_count as (
    select 'store_sales' as name, count(*) as row_count from store_sales
        union all
    select 'catalog_sales' as name, count(*) as row_count from catalog_sales
        union all
    select 'web_sales' as name, count(*) as row_count from web_sales
        union all
    select 'date_dim' as name, count(*) as row_count from date_dim
),

row_count_total as (
    select 'row_count_total' as name, sum(row_count) from row_count
)

select * from row_count
    union all
select * from row_count_total;


with store_sales_fields as (
    select
        'store_sales' as sales_type,
        ss_sold_date_sk as date_id,
        ss_net_profit as profit 
    from
        store_sales
),

catalog_sales_fields as (
    select
        'catalog_sales' as sales_type,
        cs_sold_date_sk as date_id,
        cs_net_profit as profit
    from
        catalog_sales
),

 web_sales_fields as (
    select
        'web_sales' as sales_type,
        ws_sold_date_sk as date_id,
        ws_net_profit as profit
    from
        web_sales
),
 
unioned_sales as (
    select * from store_sales_fields
        union all
    select * from catalog_sales_fields
        union all
    select * from web_sales_fields
),

date_fields as (
    select
        d_date_sk as date_id,
        d_year as year
    from
        date_dim
)

-- Group By with Sum and Count
select
    unioned_sales.sales_type,
    date_fields.year, 
    sum(abs(profit)) as sum_of_profit,
    count(*) as number_of_rows
from
    unioned_sales
    left join date_fields on unioned_sales.date_id = date_fields.date_id
where
    date_fields.year is not null
group by
    1, 2
order by 
    1,2;

alter warehouse xxlargewarehouse suspend;

The advanced part

We will use a Python, git and, docker to load data into Snowflake.

Install Python

Go to python.org, download and install Python.

On Windows you should choose Add Python to PATH

Test Python installation

On Windows: open CMD
On Mac: open Terminal

Run the following command:

python --version

Or, if that didn't work, try:

python3 --version

If you see something like this, Python is installed correctly:

Install docker desktop

Install Docker desktop from: www.docker.com/products/docker-desktop/

You may need to create an account on hub.docker.com/signup

Install git

On Mac

Type this in to the Terminal:

git --version

If git is installed, you will see the installed version number.

If git inn't installed, the installation will start.

On Windows

Install from git-scm.com/download/win choose 64-bit Git for Windows Setup.