Pivot and summarize data and create new dynamic columns - sql

Question

I have a table with the following structure
ID
Fiscal_Period
Account_ID
Opportunity_ID
Product_Type
Country
ARR
a1
FY21.12
014x4d
3xdjfr
Apple
US
5.3
a2
FY21.12
014x4d
3xdjfr
Orange
Canada
2.2
a3
FY20.5
014j3x
3xjas1
Tomato
US
1.1
a4
FY20.5
014j3x
3xjas2
Apple
Canada
4.2
I'm looking to create a version of the table that dynamically captures all the unique values in Product Type, Country, and Fiscal Period, and create custom summed ARR fields for each unique account ID:
accountID
ARR_APPLE_FY21.12
ARR_ORANGE_FY21.12
ARR_TOMATO_FY21.12
ARR_US_FY21.12
ARR_CANADA_FY21.12
ARR_TOTAL_FY21.12
ARR_APPLE_FY20.5
ARR_ORANGE_FY20.5
ARR_TOMATO_FY20.5
ARR_US_FY20.5
ARR_CANADA_FY20.5
ARR_TOTAL_FY20.5
014x4d
5.3
2.2
0
5.3
2.2
7.5
---
---
---
---
---
---
014j3x
0
0
0
0
0
0
4.2
---
1.1
1.1
4.2
5.3
I understand that this might not be the best way to organize data as it would make the dataset unnecessarily large and create a lot of fields. Primary purpose of this is to transform some data before I join it into a bigger account-level data set. I'm also open to performing this non-dynamically, as I've had some experience with Snowflake restricting my pivot abilities in the past.
I've tried using
SUM(CASE WHEN Product_Type = 'Apple' THEN ARR ELSE 0 END) AS 'ARR_Apple_FY' || SUBSTR(FISCAL_PERIOD, 3, 2) || '.' || SUBSTR(FISCAL_PERIOD, -2),
But this approach doesn't seem to work, as Snowflake doesn't like it when I use concatenates within an alias renaming. I'm ok with creating the columns statically ahead of time and grouping the sums into these columns, but would prefer to create them dynamically based on the unique values that exist within these fields.
Something like this is working, but I'm hoping for something more dynamic that also works in Snowflake.
SELECT
Account_ID,
SUM(CASE WHEN Product_Type = 'Apple' AND Fiscal_Period = 'FY21.12' THEN ARR ELSE 0 END) AS ARR_APPLE_FY2112,
SUM(CASE WHEN Product_Type = 'Orange' AND Fiscal_Period = 'FY21.12' THEN ARR ELSE 0 END) AS ARR_ORANGE_FY2112,
SUM(CASE WHEN Product_Type = 'Tomato' AND Fiscal_Period = 'FY21.12' THEN ARR ELSE 0 END) AS ARR_TOMATO_FY2112,
SUM(CASE WHEN Country = 'US' AND Fiscal_Period = 'FY21.12' THEN ARR ELSE 0 END) AS ARR_US_FY2112,
SUM(CASE WHEN Country = 'Canada' AND Fiscal_Period = 'FY21.12' THEN ARR ELSE 0 END) AS ARR_CANADA_FY2112,
SUM(CASE WHEN Fiscal_Period = 'FY21.12' THEN ARR ELSE 0 END) AS ARR_TOTAL_FY2112,
SUM(CASE WHEN Product_Type = 'Apple' AND Fiscal_Period = 'FY20.5' THEN ARR ELSE 0 END) AS ARR_APPLE_FY205,
SUM(CASE WHEN Product_Type = 'Orange' AND Fiscal_Period = 'FY20.5' THEN ARR ELSE 0 END) AS ARR_ORANGE_FY205,
SUM(CASE WHEN Product_Type = 'Tomato' AND Fiscal_Period = 'FY20.5' THEN ARR ELSE 0 END) AS ARR_TOMATO_FY205,
SUM(CASE WHEN Country = 'US' AND Fiscal_Period = 'FY20.5' THEN ARR ELSE 0 END) AS ARR_US_FY205,
SUM(CASE WHEN Country = 'Canada' AND Fiscal_Period = 'FY20.5' THEN ARR ELSE 0 END) AS ARR_CANADA_FY205,
SUM(CASE WHEN Fiscal_Period = 'FY20.5' THEN ARR ELSE 0 END) AS ARR_TOTAL_FY205
FROM myTable
GROUP BY Account_ID

Answer 1

I think you can make it dynamic if you do the following:
Pull distinct lists of Product Types and Fiscal Periods and store temp in temp tables.
Dynamically construct the SQL string you included in your question - you can do it by applying the LISTAGG() function to a cross-join of temp tables you built in step 1.
Execute the SQL stored in the string you built in Step 2 - e.g. via EXECUTE IMMEDIATE if you are using Snowflake Scripting.