Queries & Templating

DAC supports both SQL queries and semantic queries. Widgets can reference named queries or define their query shape inline.

Named SQL Queries

queries:
  total_revenue:
    sql: |
      SELECT SUM(amount) AS value FROM sales
      WHERE created_at >= '{{ filters.date_range.start }}'
        AND created_at <= '{{ filters.date_range.end }}'

  revenue_by_month:
    file: queries/revenue_by_month.sql

rows:
  - widgets:
      - name: Revenue
        type: metric
        query: total_revenue
        column: value

Named Semantic Queries

model: sales

queries:
  online_by_region:
    dimensions:
      - name: region
    metrics: [revenue]
    segments: [online]
    sort:
      - name: revenue
        direction: desc
    limit: 8

rows:
  - widgets:
      - name: Online Revenue by Region
        type: chart
        chart: bar
        query: online_by_region

Named semantic queries use the same model resolution rules as widgets. If model is omitted, DAC uses the dashboard-level model. If models aliases are defined, the query can use either a model name or an alias.

Named Query Fields

SQL Fields

Field	Type	Description
sql	string	Inline SQL
file	string	Path to a .sql file, relative to the dashboard
connection	string	Connection override

Semantic Fields

Field	Type	Description
model	string	Semantic model name. Falls back to dashboard model when omitted
dimensions	array	Dimensions to group by
metrics	string[]	Metrics to select
filters	array	Structured semantic filters
segments	string[]	Segment names to apply
sort	array	Sort instructions
limit	integer	Row limit
connection	string	Optional connection override

Semantic query fields are compiled to SQL by the backend REST API when the widget is requested.

Widget Query Sources

Every data widget needs a query. You can provide one in five ways:

1. query: reference a named query
2. sql: inline SQL
3. file: external .sql file
4. metric: semantic metric reference with a model context
5. semantic widget fields such as dimension, dimensions, metrics, filters, segments, sort, and limit

Examples:

# Named query
- name: Revenue
  type: metric
  query: total_revenue
  column: value

# Inline SQL
- name: Revenue
  type: metric
  sql: SELECT SUM(amount) AS value FROM sales
  column: value

# External file
- name: Revenue
  type: metric
  file: queries/revenue.sql
  column: value

# Semantic metric widget
- name: Revenue
  type: metric
  model: sales
  metric: revenue

# Direct semantic chart
- name: Revenue by Month
  type: chart
  chart: area
  model: sales
  dimension: created_at
  granularity: month
  metrics: [revenue]

Jinja Templating

SQL queries are processed through Jinja before execution. Semantic filter values are also templated before the backend compiles them to SQL.

Variable Interpolation

SELECT * FROM orders
WHERE region = '{{ filters.region }}'

Conditionals

SELECT * FROM orders
WHERE created_at >= '{{ filters.date_range.start }}'
  AND created_at <= '{{ filters.date_range.end }}'
{% if filters.region != 'All' %}
  AND region = '{{ filters.region }}'
{% endif %}

Date Range Filters

Date range filters expose start and end as YYYY-MM-DD strings:

WHERE created_at >= '{{ filters.date_range.start }}'
  AND created_at < DATE '{{ filters.date_range.end }}' + INTERVAL 1 DAY

The same pattern works in semantic filters:

filters:
  - dimension: created_at
    operator: between
    value:
      start: "{{ filters.date_range.start }}"
      end: "{{ filters.date_range.end }}"

Structured semantic filters target dimensions. Metric-specific predicates usually belong in the semantic model as a metric filter or as a reusable segment.

Available Variables

Variable	Source	Description
filters.<name>	Filter controls	Current filter value
filters.<date_range>.start	Date range filter	Start date
filters.<date_range>.end	Date range filter	End date

Connection Override

Any widget or named query can override the dashboard-level connection:

connection: primary_db

rows:
  - widgets:
      - name: Analytics Events
        type: table
        sql: SELECT * FROM events
        connection: analytics_db