DBT: Loading a Vertical(Key/value) table from multiple queries - snowflake-cloud-data-platform

Question

I have a use case to fill a vertical table with various product data. For internal reasons the join is messy as in our case I have only one way to link product data to the internal billing system. To get around this I am trying to feed multiple queries into a vertical table. I am aware of using an incremental configuration to populate a query but I have not found anything in the documentation to support loading the result of multiple queries into a more narrow/vertical table.
The table would be structured like the following:
UNIQUEID
ACCOUNTID
PRODUCTID
KEY
VALUE
PPD
DATE
asdfasd
099080as
8998asfas
500.00
10
1.5
6/28/2022
EDIT: Adding some more clarification as a numerical key/value seems off I know. But in this context a given account id can have only one product id and only one target/value for each productid
Through snowflake tasks I would have several queries like:
INSERT INTO TABLE (fields...)
SELECT
UNIQUEID,
ACCOUNTID,
PRODUCTID,
TARGET,
VALUE,
PPD,
CURRENT_DATE() as date
FROM accounttable t
INNER JOIN producttable t2 ON
t.accountid = t2.accountid
where t2.productid = '5898988asdfas'
and t.type = 'prodtype'
Without getting too much into the data structure problem I have I can only link to the value I need by the account Id but need to specify the productId in each query loading into it so I can ensure that I get the right product and account data each time.
Just looking for how to use DBT to do something like this if it is even possible at all.

Answer 1

Assuming that your Product IDs (like 5898988asdfas) are not present in your database, or not otherwise related to the account IDs the way you need them to be, you have a couple of options:
Create a CSV that maps account ID's to product ID's (in Excel or Sheets) and upload this table to your database using dbt seed. Then you would use that mapping to join the other tables together, like:
FROM accounttable t
INNER JOIN mappingtable on t.accountid = mappingtable.accountid
INNER JOIN producttable t2
ON t.accountid = t2.accountid
AND mappingtable.productid = t2.productid
where t.type = 'prodtype'
Alternatively, you could store the values of productid in a Jinja list and build a select statement using a union
{% set ids = ["5898988asdfas", "6898988asdfas", "7898988asdfas"] %}
{% for id in ids %}
SELECT
UNIQUEID,
ACCOUNTID,
PRODUCTID,
TARGET,
VALUE,
PPD,
CURRENT_DATE() as date
FROM accounttable t
INNER JOIN producttable t2 ON
t.accountid = t2.accountid
where t2.productid = '{{ id }}'
and t.type = 'prodtype'
{% if not loop.last %} union all {% endif %}
{% endfor %}
Either of the above solutions are a single select statement, so they can be a single model in your dbt project. You can materialize them however you like, but probably not incrementally (incremental models are really intended for something like an event stream, where new immutable data is always being appended to the source dataset).