Welcome to the Powered by Tasty Bytes - Zero to Snowflake Quickstart focused on Transformation!
Within this Quickstart we will walk through a large set of Snowflake functionality covering key features like Zero Copy Cloning and Time-Travel to deliver on a Tasty Bytes business requirement.
Within this Quickstart we will follow a Tasty Bytes themed story via a Snowsight SQL Worksheet with this page serving as a side by side guide complete with additional commentary, images and documentation links.
This section will walk you through logging into Snowflake, Creating a New Worksheet, Renaming the Worksheet, Copying SQL from GitHub, and Pasting the SQL we will be leveraging within this Quickstart.
As part of Tasty Bytes Fleet Analysis, our Developer has been tasked with creating and updating a new Truck Type column within the Raw layer Truck table that combines the Year, Make and Model together.
Within this step, we will first walk through standing up a Development environment using Snowflake Zero Copy Cloning for this development to be completed and tested within before rolling into production.
Thanks to Snowflake's unique architecture, we can instantly create a snapshot of our production raw_pos.truck
using CLONE functionality and name it raw_pos.truck_dev
.
Let's now run our next set our queries to set our tb_dev
Role and tb_101
Warehouse context and create the table clone; noting here that we do not need to set Warehouse context since cloning does not require one. This query will provide a Table TRUCK_DEV successfully created
result.
USE ROLE tb_dev;
USE DATABASE tb_101;
CREATE OR REPLACE TABLE raw_pos.truck_dev CLONE raw_pos.truck;
With our Zero Copy Clone instantly available we can now begin to develop against it without any fear of impacting production. However, before we make any changes let's first run some simple queries against it and test out Snowflake's Result Set Cache.
Now that we are going to query our Table, we will need to use our tb_dev_wh
Warehouse.
Let's kick off the next two queries with the second statement producing an result set consisting of our trucks, their years, make and models while making sure we ORDER BY our truck_id
Column.
USE WAREHOUSE tb_dev_wh;
SELECT
t.truck_id,
t.year,
t.make,
t.model
FROM raw_pos.truck_dev t
ORDER BY t.truck_id;
To test Snowflake's Result Set Cache, the next query we run will be identical to what we just ran.
However, we can now take things a step further and access the Query Profile showcasing this query returned results instantly as the the results came from our Result Set Cache.
SELECT
t.truck_id,
t.year,
t.make,
t.model, --> Snowflake supports Trailing Comma's in SELECT clauses
FROM raw_pos.truck_dev t
ORDER BY t.truck_id;
Within this step, we will now will Add and Update a Truck Type column to the Development Truck Table we created previously while also addressing the typo in the Make
field.
To begin this section, let's make sure we correct the typo by executing our next query which leverages UPDATE to change rows in our truck_dev
WHERE the make is equal to Ford_.
This query will provide a Number of Rows updated
result set.
UPDATE raw_pos.truck_dev
SET make = 'Ford' WHERE make = 'Ford_';
With the typo handled, we can now build the query to concatenate columns together that will make up our Truck Type. Please execute the next query where we will see CONCAT and REPLACE leveraged.
SELECT
truck_id,
year,
make,
model,
CONCAT(year,' ',make,' ',REPLACE(model,' ','_')) AS truck_type
FROM raw_pos.truck_dev;
To start, please execute the next query which uses ALTER TABLE... ADD COLUMN to create an empty truck_type
column of Data Type VARCHAR to our truck_dev
table.
This query will provide a Statement executed successfully
result.
ALTER TABLE raw_pos.truck_dev
ADD COLUMN truck_type VARCHAR(100);
With the column in place, we can kick off the next query which will UPDATE the new, empty truck_type
column using the Truck Type concatenation we built in the previous section.
This query will provide a Number of Rows Updated
result set.
UPDATE raw_pos.truck_dev
SET truck_type = CONCAT(year,make,' ',REPLACE(model,' ','_'));
After successfully updating the data, let's now run a quick query against the table to see how things look in our truck_type
column.
SELECT
truck_id,
year,
truck_type
FROM raw_pos.truck_dev
ORDER BY truck_id;
Uh oh! Thank goodness we were smart developers and didn't do this sort of thing blindly in production.
It looks like we messed up the truck_type
concatenation. We will need to resolve this in our next section.
Althoug we made a mistake on the Update statement earlier and missed adding a space between Year and Make. Thankfully, we can use Time Travel to revert our table back to the state it was after we fixed the misspelling so we can correct our work.
To start our recovery process, kick off the next query which will use the Snowflake QUERY_HISTORY function to retrieve a list of all update statements we have made against our truck_dev
Table.
SELECT
query_id,
query_text,
user_name,
query_type,
start_time
FROM TABLE(information_schema.query_history())
WHERE 1=1
AND query_type = 'UPDATE'
AND query_text LIKE '%raw_pos.truck_dev%'
ORDER BY start_time DESC;
As expected, we see our typo correction as well as our update and their associated unique query_id's. Please run the next query which creates a query_id
SQL Variable that we will use to revert our changes via Time-Travel in the next step.
After execution you will recieve a Statement executed successfully
result.
SET query_id =
(
SELECT TOP 1
query_id
FROM TABLE(information_schema.query_history())
WHERE 1=1
AND query_type = 'UPDATE'
AND query_text LIKE '%SET truck_type =%'
ORDER BY start_time DESC
);
With our bad query_id stored as a Variable, we can execute the next query which will replace our truck_dev
Table with what it looked like BEFORE the incorrect query_id statement using Time-Travel.
SELECT
truck_id,
make,
truck_type
FROM raw_pos.truck_dev
BEFORE(STATEMENT => $query_id)
ORDER BY truck_id;
Please refer to the list below for the other Time-Travel Statement options available.
Happy with our results, let's now execute the next query to recreate the table. This query will provide a Table TRUCK_DEV successfully created.
result.
CREATE OR REPLACE TABLE raw_pos.truck_dev
AS
SELECT * FROM raw_pos.truck_dev
BEFORE(STATEMENT => $query_id); -- revert to before a specified Query ID ran
To conclude, let's run the correct update statement which will provide a Number of Rows Updated
result set.
UPDATE raw_pos.truck_dev t
SET truck_type = CONCAT(t.year,' ',t.make,' ',REPLACE(t.model,' ','_'));
Based on our previous efforts, we have addressed the requirements we were given and to complete our task need to push our Development into Production.
Within this step, we will swap our Development Truck table truck_dev
with what is currently available in Production.
Please kick off the next two queries where we first assume the more privileged accountadmin
role. As a accountadmin
the second query utilizes ALTER TABLE... SWAP WITH to promote our truck_dev
table to truck
and vice versa.
Once complete you will recieve a Statement executed successfully.
result.
USE ROLE accountadmin;
ALTER TABLE raw_pos.truck_dev
SWAP WITH raw_pos.truck;
To confirm our process was successful, let's now take a look at the Production truck
table so we can validate the swap was successful and the truck_type
results are valid.
SELECT
t.truck_id,
t.truck_type
FROM raw_pos.truck t
WHERE t.make = 'Ford';
We can officially say our developer has completed their assigned task. With the truck_type
column in place and correctly calculated, our sysadmin
can clean up the left over Table and sign off for the day.
To remove the Table from our Database, please execute the next query which leverages DROP TABLE. This query will provide a TRUCK successfully dropped.
result.
DROP TABLE raw_pos.truck;
Uh oh!! That result set shows that even our accountadmin
can make mistakes. We incorrectly dropped production truck
and not development truck_dev
! Thankfully, Snowflake's Time-Travel can come to the rescue again.
Hurry up and run the next query before any systems are impacted which will UNDROP the truck
table. This query will provide a Table TRUCK successfully restored.
result.
UNDROP TABLE raw_pos.truck;
Alright, now let's officially close things out by running the final query to correctly drop truck_dev
. This query will provide a TRUC_DEV successfully dropped.
result.
DROP TABLE raw_pos.truck_dev;
Fantastic work! You have successfully completed the Tasty Bytes - Zero to Snowflake - Transformation Quickstart.
By doing so you have now:
If you would like to re-run this Quickstart please leverage the Reset scripts in the bottom of your associated Worksheet.
To continue your journey in the Snowflake AI Data Cloud, please now visit the link below to see all other Powered by Tasty Bytes - Quickstarts available to you.