Warehouse shares

If your analysts live in the warehouse, do not make them call an API. The share puts the entity graph in your account as tables you can join to directly.

What arrives

Shares are read-only, versioned, and refreshed on the same cadence as the API — hourly for Tier 1 jurisdictions. No copy, no egress cost, no pipeline of your own to keep alive.

Joining to your own data

Resolve once through the API, store entity_id on your account table, and every warehouse join afterwards is a key join rather than a fuzzy one.

select
  a.account_id,
  a.name              as crm_name,
  e.name              as register_name,
  e.status,
  e.nace_primary,
  g.commune,
  h.value             as headcount
from crm.accounts a
join spotit.entities e on e.entity_id = a.spotit_entity_id
left join spotit.geo g on g.geo_id = e.geo_id
left join spotit.entity_attributes h
  on h.entity_id = e.entity_id and h.attribute = 'headcount'
where e.status = 'active'

Incremental models

entity_history and events are append-only and carry a monotonic ingested_at. That is the only watermark an incremental model needs.

select *
from spotit.events
where ingested_at > (select coalesce(max(ingested_at), '2009-01-01') from this_model)

Platform notes

• Snowflake — a native Secure Data Share. Accept it and the database appears; no warehouse of ours runs in your account.
• BigQuery — Analytics Hub listing, subscribed into a linked dataset in your own project and region.
• Databricks — Delta Sharing, so the same share works from a notebook or from an external Delta reader.
• All three are available on Business and Enterprise. The EU deployment is the default; ask before you pin one elsewhere.

Cost

The share is billed as a flat add-on rather than per credit, because a warehouse join is not a metered call. Reads inside your own account cost you whatever your warehouse charges and cost us nothing. See pricing.