> ## Documentation Index > Fetch the complete documentation index at: https://help.treble.ai/llms.txt > Use this file to discover all available pages before exploring further. # Dimensions > Catalog tables for enriching your fact queries: agents, teams, templates, flows, channels, and flow nodes. # Dimensions Dimensions are small catalog tables that describe the entities your facts reference. Fact tables already carry the most common names inline (`agent_name`, `poll_name`, `team_name`…), so you only need a dimension join when you want attributes beyond the name — an agent's email, an HSM's approval status, a node's AI configuration. ## dim\_agents One row per agent account. | Column | Type | Description | | ------------ | ------ | ------------------------------------------------- | | `id` | Int32 | Agent identifier (joins `agent_id` in facts) | | `company_id` | Int32 | Your company (filtered automatically) | | `first_name` | String | First name | | `last_name` | String | Last name | | `email` | String | Agent's email | | `active` | Bool | Whether the agent is currently available (online) | | `agent_type` | String | Role: `AGENT`, `ADMIN`, or `SUPERVISOR` | ## dim\_teams One row per team. | Column | Type | Description | | ------------ | ------ | --------------- | | `id` | Int32 | Team identifier | | `company_id` | Int32 | Your company | | `name` | String | Team name | ## dim\_tags One row per tag (queue/group inside a team). | Column | Type | Description | | ------------ | ------ | ----------------------- | | `id` | Int32 | Tag identifier | | `company_id` | Int32 | Your company | | `name` | String | Tag name | | `team_id` | Int32 | Team the tag belongs to | ## dim\_agent\_tags One row per agent–tag assignment — which agents work which queues. | Column | Type | Description | | ------------ | ------ | ---------------- | | `agent_id` | Int32 | The agent | | `company_id` | Int32 | Your company | | `tag_id` | Int32 | The tag assigned | | `tag_name` | String | Tag name | | `team_id` | Int32 | Team of the tag | | `team_name` | String | Team name | ## dim\_hsms One row per HSM (WhatsApp template). | Column | Type | Description | | --------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `id` | Int32 | Template identifier (joins `hsm_id` in facts) | | `company_id` | Int32 | Your company | | `name` | String | Template name | | `status` | String | Lifecycle status: `APPROVED`, `PROCESSING`, `REJECTED`, `PAUSED`, `DISABLED`, `ARCHIVED`, `PENDING_DELETION` | | `category` | String | Meta category: `MARKETING`, `UTILITY`, `AUTHENTICATION`, plus legacy categories on older templates (`ACCOUNT_UPDATE`, `ALERT_UPDATE`, `TRANSACTIONAL`, `OTP`, …) | | `template_type` | String | `TEXT`, `RICH`, or `MEDIA` | | `content` | String | Template text content | | `provider_id` | String | Meta's template id | ## dim\_polls One row per flow **version**. When you edit a flow in the editor, a new version (new `poll_id`) is created; `original_poll_id` groups all versions of the same logical flow. | Column | Type | Description | | ------------------ | ------ | -------------------------------------------------- | | `poll_id` | Int32 | Flow version identifier (joins `poll_id` in facts) | | `poll_name` | String | Flow name | | `original_poll_id` | Int32 | Identifier of the flow across versions | | `version` | UInt64 | Version number | | `hidden` | Bool | Whether this version is hidden in the platform | | `company_id` | Int32 | Your company | To analyze a flow across edits, group facts by `original_poll_id` via this dimension rather than by raw `poll_id`. ## dim\_channels One row per WhatsApp line. | Column | Type | Description | | --------------- | ---------- | ------------------------------------------------ | | `company_id` | Int32 | Your company | | `channel_id` | Int32 | Channel identifier (joins `channel_id` in facts) | | `whatsapp_line` | String | The line's phone number | | `display_name` | String | The line's display name | | `created_at` | DateTime64 | When the line was created | ## dim\_poll\_nodes One row per node of each flow version — the structure behind [`fact_treble_session_nodes`](/en/docs/data-warehouse-v2/fact-treble-session-nodes). For AI nodes it includes the full AI agent configuration. | Column | Type | Description | | --------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------- | | `node_id` | Int32 | Node identifier (joins `node_id` / `last_node_id` in facts) | | `company_id` | Int32 | Your company | | `poll_id` | Int32 | Flow version the node belongs to | | `node_type` | String | Node type: `QUESTION`, `AI_TASK_AUTOMATOR`, `AI_AGENT`, `WHATSAPP_FLOW`, `META_EVENTS`, `HELPDESK_INTEGRATION`, and others | | `question_id` | Int32 | Question identifier, when the node asks one; `0` otherwise | | `question_text` | String | The question's text | | `question_type` | String | `OPEN` (free text), `CLOSED` (predefined options), or `AGENT` (hands over to an agent); empty for non-question nodes | | `hsm_id` | Int32 | HSM template the node sends, if any; `0` otherwise | | `ai_automator_id` | Int64 | AI agent configured on the node, if any; `0` otherwise | | `ai_automator_name` | String | AI agent's name | | `ai_automator_description` | String | AI agent's description | | `ai_automator_version` | Int32 | Latest version of the AI agent linked to this node | | `ai_automator_active` | Bool | Whether that latest version is active | | `ai_automator_tools` | String | Tools enabled for the AI agent (JSON) | | `ai_automator_instructions` | String | The AI agent's instructions | ## Example: enrich a funnel with node text ```sql theme={null} SELECT f.node_id, n.node_type, n.question_text, uniqExact(f.session_id) AS conversations_reached FROM fact_treble_session_nodes AS f LEFT JOIN dim_poll_nodes AS n ON n.company_id = f.company_id AND n.poll_id = f.poll_id AND n.node_id = f.node_id WHERE f.poll_id = {your_poll_id} AND f.entered_at >= today() - 30 GROUP BY f.node_id, n.node_type, n.question_text ORDER BY conversations_reached DESC ```