Event Warehouse

Write collected events to your data warehouse for long-term storage and analysis.

Overview

When a forwarding rule targets a warehouse source (BigQuery, Snowflake, Databricks), events are batched, serialized, uploaded to staging, and bulk-loaded into per-event-type tables. Every event carries its originating source identity, consent state, and is partitioned on its event timestamp for efficient time-range queries.

Configuration

Navigate to Streams > Forwarding
Click Create Forwarding Rule
Select the event source and choose a warehouse source as the target
Configure the warehouse settings:
- Schema name (default: signalsmith_events) — target dataset/schema
- Custom table name (optional) — override the auto-derived table name
- Partition column (default: timestamp) — column used for day-partitioning/clustering
Click Save

Table Naming

Zeotap creates one table per event type. The table name depends on whether the forwarding rule is scoped to a specific source and whether a custom table name is configured.

Event Type	Source Prefix	Table Name
`track`	none	`tracks`
`identify`	none	`identifies`
`page`	none	`pages`
`screen`	none	`screens`
`group`	none	`groups`
`track`	”Web App”	`web_app_tracks`
`identify`	”Mobile App”	`mobile_app_identifies`
`purchase` (custom)	none	`purchases`
`app.install` (custom)	none	`app_installs`

Source prefixes are derived from the event source (write key) name: lowercased, non-alphanumeric characters replaced with underscores, and truncated to a safe length. Custom event types are pluralized idempotently (names already ending in s are left alone).

To override the auto-naming, set Custom Table Name on the forwarding rule. This is useful when downstream consumers expect a specific table name.

Table Schema

Every event table has the same columns regardless of warehouse dialect:

Column	Type	Description
`id`	string	Message ID (unique per event)
`workspace_id`	string	Workspace identifier
`write_key_id`	string	Which event source sent this event
`write_key_name`	string	Human-readable source name for quick inspection
`user_id`	string	Identified user ID
`anonymous_id`	string	Anonymous user ID
`event`	string	Event name (for track)
`properties`	JSON / VARIANT	Event properties
`traits`	JSON / VARIANT	User traits (for identify)
`context`	JSON / VARIANT	Event context (page, device, campaign, etc.)
`timestamp`	timestamp	When the event occurred (client-provided)
`received_at`	timestamp	When Zeotap received the event
`sent_at`	timestamp	When the client sent the event
`group_id`	string	Group ID (for group events)
`name`	string	Page/screen name
`category`	string	Page/screen category
`server_ip`	string	Client IP address
`consent_string`	string	IAB TCF v2 consent string
`us_privacy`	string	IAB CCPA US Privacy string (e.g., `1YNN`)
`consent_categories`	JSON / VARIANT	Granted consent categories as a JSON object

Partitioning and Clustering

New event tables are partitioned (or clustered, depending on the warehouse) on the timestamp column so time-range queries scan only the relevant partitions:

Warehouse	Strategy
BigQuery	Day-partitioned on `timestamp` via `TimePartitioning`
Snowflake	`CLUSTER BY (timestamp)` after table creation
Databricks	`CLUSTER BY (timestamp)` on Delta tables

You can override the partition/clustering column per forwarding rule by setting Partition Column in the rule config. This is useful when you have a custom event timestamp field that better reflects your query patterns.

Note: BigQuery cannot retroactively partition existing tables. If you need to change the partition column or enable partitioning on a pre-existing table, you must drop and recreate it.

Every event row carries structured consent context extracted from context.consent:

consent_string — IAB TCF v2 consent string or your own consent string
us_privacy — IAB CCPA US Privacy string
consent_categories — JSON object of granted categories (e.g., {"analytics": true, "marketing": false})

This lets you query consent-aware cohorts directly in SQL:


-- Count users who opted out of marketing
SELECT COUNT(DISTINCT user_id)
FROM web_app_tracks
WHERE JSON_EXTRACT(consent_categories, '$.marketing') = false
  AND timestamp >= '2026-01-01';

Schema Evolution

The column list is fixed — new event properties are stored inside the properties JSON column
Query JSON columns using your warehouse’s JSON functions (JSON_EXTRACT, ->, ::variant)
Custom event types automatically get their own table on first event arrival

API Reference

See Events API for warehouse configuration endpoints.

Next Steps

Event forwarding — Configure routing rules
Consent management — Set up consent categories
Sending events — Include consent in event payloads