Dataform Support

Querylab.io natively supports Google Cloud Dataform SQLX templates. Write and test Dataform models directly in the editor with completions, diagnostics, and mapping management — no Dataform CLI required.

---

Getting Started

Dataform mode activates automatically when Querylab.io detects SQLX patterns in your SQL:

• ${ ... } template expressions
• config { ... } blocks
• js { ... } blocks
• pre_operations { ... } or post_operations { ... } blocks

Paste any Dataform SQLX into the editor and the Dataform panel appears automatically.

Enable/Disable

Dataform mode is enabled by default. To toggle it:

Settings > SQL Frameworks > Enable Dataform Mode

See Settings for all Dataform-related settings.

---

The Dataform Panel

When Dataform mode activates, a Dataform button appears in the status bar at the bottom of the editor. Click it to open the Dataform panel on the right side.

• Left-click the status bar button to toggle the panel
• Right-click the button to deactivate Dataform mode for this tab

The panel has six sections:

Active Directives

Lists all Dataform directives found in your SQL:

• ref('table_name') references
• resolve('project.dataset.table') references
• dataform.projectConfig.vars.name variable references

Each shows the directive and its resolution status. Click unresolved items to add a mapping.

Detected Tables

Tables from your FROM and JOIN clauses that are not yet converted to Dataform refs. Click Convert to Dataform ref to replace the table reference with ${ ref("table_name") }.

Detected Values

String literals that could be extracted as Dataform variables. Click Convert to Dataform variable to replace the value with ${ dataform.projectConfig.vars.name }.

Table Mappings

Manage your ref() and resolve() mappings. Each mapping connects a Dataform reference to its BigQuery table.

• ref mappings resolve ${ ref("name") } to a BigQuery table
• resolve mappings resolve ${ resolve("name") } to a fully-qualified table

Add, search, and remove mappings the same way as dbt mappings. Mappings are saved globally.

Variable Mappings

Manage your project variable mappings. Each maps a variable name to its resolved value for ${ dataform.projectConfig.vars.name } expressions.

Options

Option	Description
Incremental mode	When enabled, when(incremental()) blocks are kept during preprocessing. When disabled, they are stripped.
Auto-apply saved mappings	Automatically resolve ref() and vars using your saved mappings.

---

SQLX Syntax Reference

Dataform uses SQLX template syntax, which differs from dbt's Jinja:

config {
  type: "view",
  schema: "analytics"
}

SELECT
  id,
  name,
  created_at
FROM ${ ref("source_table") }
WHERE created_at >= ${ dataform.projectConfig.vars.start_date }

Common Patterns

Pattern	Purpose
${ ref("table") }	Reference a Dataform model
${ resolve("project.dataset.table") }	Reference any BigQuery table
${ self() }	Reference current model's output table
${ dataform.projectConfig.vars.name }	Access a project variable
${ when(incremental(), "WHERE id > last_id") }	Conditional SQL for incremental queries

Config Blocks

SQLX files can include configuration blocks at the top:

config {
  type: "incremental",
  schema: "analytics",
  bigquery: {
    partitionBy: "DATE(created_at)",
    clusterBy: ["user_id"]
  }
}

---

Inlay Hints

When enabled, resolved values appear inline after SQLX expressions as faded annotations:

SELECT *
FROM ${ ref("users") }         -- → my_project.analytics.users
WHERE date >= ${ dataform.projectConfig.vars.start_date }  -- → 2024-01-01

Toggle in Settings > SQL Frameworks > Dataform > Inlay Hints.

---

Diagnostics

Querylab.io flags Dataform issues with editor squiggles:

• Unresolved ref: ${ ref("unknown") } — warning with quick fix to add mapping
• Unresolved resolve: ${ resolve("unknown") } — warning
• Unresolved variable: ${ dataform.projectConfig.vars.unknown } — warning

Click the lightbulb or press Cmd/Ctrl+. to access quick fixes.

---

Import Config

Bulk-import mappings from your Dataform project's dataform.json:

1. Click the Import button (upload icon) in the panel header
2. Paste the contents of your dataform.json file
3. Review the preview showing counts of tables and variables
4. Choose Merge (add to existing) or Replace (overwrite all)

---

Context Menu

Right-click in the editor when Dataform mode is active for additional actions:

• Convert to Dataform ref — Replace selected table name with ${ ref("...") }
• Convert to Dataform variable — Replace selected value with a project variable

---

dbt vs Dataform

Feature	dbt	Dataform
Template syntax	Jinja: {{ }}, {% %}	SQLX: ${ }
Ref syntax	{{ ref('model') }}	${ ref("model") }
Source syntax	{{ source('schema', 'table') }}	${ resolve("project.dataset.table") }
Variables	{{ var('name') }}	${ dataform.projectConfig.vars.name }
Incremental	{% if is_incremental() %}	${ when(incremental(), "sql") }
Import from	manifest.json	dataform.json
Export	Generate model YAML	Not supported

Both modes can be enabled simultaneously — Querylab.io auto-detects which syntax is in use per tab.

---

Related Features

• dbt Support — dbt Jinja template support
• Settings — dbt and Dataform settings
• Code Quality — Diagnostic rules