# Building queries

> Pick dimensions and measures, add filters, sorts, and a limit, and watch results update as Flax recompiles and re-runs the query in your warehouse.

*[View this page in the Flax docs](https://flax-analytics.com/docs/exploring/building-queries)*

A query in Explore is a set of choices: what to group by, what to measure, which
rows to include, and how to order them. This page walks through each control and
how results are produced.

## Choose dimensions and measures

From the field-picker, toggle on:

- **Dimensions** — attributes you group or slice by, such as `region` or
  `order_date`. Time dimensions add a **Grain** control (day, week, month,
  quarter, year) so you can roll dates up to the level you want.
- **Measures** — aggregated values, such as `revenue` or `order_count`.

Flax groups by every selected dimension and aggregates every selected measure. You
need at least one dimension or measure before you can run the query. Fields from a
joined model appear in their own section, and Flax resolves the join for you when
you combine fields across related models.

## Add filters

Filters narrow the rows the query considers. Add a filter on a dimension or measure
and pick an operator: equals or not-equals, is one of / not one of a set, the
numeric and date comparisons (greater/less than, or between), contains for text,
and is-null / is-not-null. All filter values are passed as bound parameters, never
concatenated into SQL.

> [!NOTE]
> Filters you apply are always layered on top of any
> [row-access filters](/docs/modeling/row-access-filters) defined in the model.
> You can never widen your access by removing a filter.

## Sort, compare, and calculate

- **Sort by** a single field, ascending or descending (toggle **Desc**).
- **Compare** to the previous period with the **vs. previous period** switch,
  available once you have a time-grain dimension and a measure selected.
- **Table calculations** add a **running total** or **% of total** to a measure.

## Running the query

Building a query is deliberate: your selections do not run until you click **Run**.
Flax then compiles the SQL for your connection's dialect (Postgres, Snowflake,
Databricks, or BigQuery) and executes it in your warehouse, and the result view
refreshes.

Large results are capped server-side (up to 5,000 rows) so the browser stays
responsive; when a result is cut off, Flax flags it as truncated. Reach for tighter
filters or a coarser grain if you hit the cap.

## From result to output

Once a result looks right you can:

- Switch it to a [chart](/docs/visualizations/chart-types) and save it.
- Open it in a [sheet](/docs/sheets/overview) for formulas and formatting.
- [Drill down](/docs/exploring/drill-down) into a value to see its rows.
- Inspect the [compiled SQL](/docs/exploring/query-detail).

## Next steps

- [The SQL runner](/docs/exploring/sql-runner)
- [Drill-down](/docs/exploring/drill-down)
- [Query detail](/docs/exploring/query-detail)