PB

How to Use

After installing PB and configuring your connections, you need to change inputs.yaml with names of your source tables. Once that is done, please mention their names as edge_sources in profiles.yaml and define specs for creating ID stitcher / feature table.

Use this command to validate that your project shall be able to access the warehouse specified in connections and create material objects there.

pb validate access

You can use this command to generate SQL, which will also tell you in case there are syntax errors in your model YAML file.

pb compile

If there are no errors, use this command to create the output table on the warehouse.

pb run

SQL queries for data analysis.

Let's assume that the materialized table created by PB was named MATERIAL_USER_STITCHING_26f16d24_29 , inside schema RUDDER_360 of database RUDDER_EVENTS_PRODUCTION. The materialized table name will change with each run, the view USER_STITCHING will point to the most recently created one.

Total number of records:

select count(*) from RUDDER_EVENTS_PRODUCTION.RUDDER_360.USER_STITCHING;

Total number of distinct records (rudder_id):

select count(distinct rudder_id) from RUDDER_EVENTS_PRODUCTION.RUDDER_360.USER_STITCHING;

Max mappings to a single canonical ID:

select rudder_id, count(other_id) as "CNT"
from RUDDER_EVENTS_PRODUCTION.RUDDER_360.USER_STITCHING
group by rudder_id
order by CNT DESC;

Say there was a canonical ID '0013d4fa-fdf7-5736-85d1-063378251398' that had more than 1000 mappings. So to check more on other ID types and their count:

select count (distinct other_id) as "OTHER_ID_COUNT", other_id_type from RUDDER_EVENTS_PRODUCTION.RUDDER_360.USER_STITCHING
where rudder_id = '0013d4fa-fdf7-5736-85d1-063378251398'
group by other_id_type;

Know More

See public docs for more information on using PB.