There are total of 198M records, determine and validate algorithm for identifying primary record in Tally Snapshots for each non-payg products. 

• Each Non-PAYG Products (RHEL for x86, OpenShift): Have 4 base dimension combinations (sla × usage), with billing_provider and billing_account_id fields set to _ANY. 
• Validate above point and check what permutations exists and how the data looks like for primary record. 
• Check for edge cases like any of the value empty.

AC:

1. Identify which of the 4 snapshots is the primary record for non-payg product in prod.
2. Write rules if there are any edge cases and document them for non-payg
3. Document queries for finding primary record and any edge case in Design doc https://docs.google.com/document/d/1_lcBwWnR9upeX2V0QDp5Mx_yo_b1aZGlsjNjprbQD6E/edit?usp=sharing
4. Add a summary to the document showing, per product, the number of records with _ANY values and the number of primary records without _ANY. This will help prioritize larger products when setting is_primary=true via the REST API.