Bring Obsidian Dataview query power to VS Code's built-in Markdown preview.

Query your workspace's Markdown notes using DQL (Dataview Query Language) — with support for TABLE, LIST, TASK, and CALENDAR views, plus inline expressions and DataviewJS.

Features

🔍 Automatic indexing of all Markdown files with live-updating
📊 TABLE queries with custom columns, sorting, GROUP BY, FLATTEN, and aggregate functions (sum/average/min/max/length(rows)) with automatic summary rows
📋 LIST queries with optional extra fields
✅ TASK queries across all files, with due dates and emoji shorthands (🗓️ ✅ ⏳ 🛫)
📅 CALENDAR view for date-based data
🧮 Inline expressions — `= this.file.name` renders current page metadata
🏷️ Inline fields — [key:: value] and key:: value syntax, with visual rendering
🧩 DataviewJS support (optional, with trusted-path security control)
📦 70+ built-in functions — date, string, list, math, logic operations
🏷️ Full metadata support: YAML frontmatter, tags, wikilinks, implicit fields
⚡ Parallel indexing — 8x concurrent file processing for large workspaces

Usage

DQL Queries

Create a ```dataview code block in any Markdown file:

```dataview
TABLE file.ctime AS "Created", file.mtime AS "Modified"
FROM "notes"
WHERE file.size > 100
SORT file.mtime DESC
LIMIT 10
```

Query Types

Type	Description	Example
`TABLE`	Tabular output with configurable columns	`TABLE field1, field2 AS "Label" FROM #tag`
`LIST`	Bulleted list of files	`LIST FROM "projects"`
`TASK`	Task list across files	`TASK FROM #todo`
`CALENDAR`	Calendar view by date	`CALENDAR due FROM #project`

Pipeline Commands

TABLE column1, column2 AS "Name"
FROM #tag AND "folder"
WHERE condition
SORT field ASC
GROUP BY category
FLATTEN tags
LIMIT 20

FROM — filter by tag (#tag), folder ("path"), file, or link ([[page]])
WHERE — filter rows with boolean expressions
SORT — sort by one or more fields (ASC/DESC)
GROUP BY — group results with rows.field swizzling
FLATTEN — expand array fields into multiple rows
LIMIT — restrict result count

Aggregate Queries

Aggregate functions collapse the matched rows into a single summary row (or per-group rows with GROUP BY). Multi-line column lists are supported:

```dataview
TABLE WITHOUT ID
  sum(file.tasks.length) AS "Total Tasks",
  average(budget) AS "Avg Budget",
  min(budget) AS "Min Budget",
  max(budget) AS "Max Budget",
  length(rows) AS "Projects"
FROM #project
```

Inline Expressions

Use backtick-wrapped expressions in your prose:

Created: `= this.file.ctime`
Tags: `= this.file.tags`
Rating: `= this.rating`

Metadata Sources

Dataview indexes three sources of metadata:

Implicit fields — file.name, file.path, file.ctime, file.mtime, file.size, file.tags, file.inlinks, file.outlinks, file.tasks, and more
YAML Frontmatter — All fields in --- blocks are queryable with dot notation for nested objects
Inline fields — Key:: Value syntax within Markdown body

Available Functions

Category	Functions
Constructors	`date()`, `dur()`
Type checking	`typeof()`, `isnumber()`, `isstring()`, `isdate()`, `islink()`, etc.
Numeric	`round()`, `floor()`, `ceil()`, `min()`, `max()`, `sum()`, `average()`
String	`lower()`, `upper()`, `contains()`, `length()`, `regexmatch()`, `replace()`, `split()`, `join()`, `trim()`
List	`filter()`, `map()`, `flat()`, `extract()`, `sort()`, `reverse()`, `nonnull()`
Date	`dateformat()`, `dateadd()`, `datestart()`, `dateend()`, `striptime()`
Logic	`default()`, `choice()`

DataviewJS (Optional)

Enable DataviewJS in settings (dataview.enableJavascript: true), then use:

```dataviewjs
dv.table(["Name", "Created"], dv.pages("#project")
  .sort(p => p.file.ctime, "desc")
  .limit(10)
  .map(p => [p.file.name, p.file.ctime]))
```

Use dv.span() / dv.paragraph() to append formatted text (inline markdown such as **bold** is rendered):

```dataviewjs
const projects = dv.pages("#project");
dv.span("**Projects:** " + projects.length);
dv.paragraph("Total budget: " + projects.array().reduce((s, p) => s + (p.budget || 0), 0));
```

⚠️ Security Warning: DataviewJS runs arbitrary JavaScript from your notes in the Markdown preview. Only enable if you trust all content in your workspace.

Configuration

Setting	Default	Description
`dataview.enableJavascript`	`false`	Enable DataviewJS code blocks
`dataview.enableInlineQueries`	`true`	Enable inline `= expr` expressions
`dataview.defaultDateFormat`	`"yyyy-MM-dd"`	Default date format
`dataview.warnOnEmptyResult`	`true`	Show warning for empty query results
`dataview.taskEmojis`	`true`	Parse task emoji shorthands (🗓️, ✅, etc.)
`dataview.excludedFolders`	`["node_modules", ".git", ".vscode"]`	Folders to exclude from indexing
`dataview.javascriptTrustedPaths`	`[]`	Folder paths trusted for DataviewJS. Empty = trust all.

Commands

Dataview: Re-index workspace — Force a full reindex of all Markdown files
Dataview: Show index statistics — Display page/tag/inlink counts

Requirements

VS Code 1.95.0 or later

Known Limitations

DataviewJS is sandboxed in the preview webview and cannot access the filesystem or VS Code API
TASK queries are read-only (checking a task in the preview does not modify the source file)
Link dereferencing ([[Page]].field) is limited to indexed pages
Very large workspaces (10,000+ files) may experience slower indexing on startup

Development

npm test          # type-check + run the automated test suite
npm run build     # type-check + bundle extension & preview
npm run package   # create the .vsix

The suite covers the parser, expression engine, query engine (integration against the real test/ vault), and markdown-it plugin rendering. Standalone debug harnesses (scripts/debug-query.ts, scripts/debug-links.ts) run the real parser/engine without launching VS Code.

License

MIT

Dataview — Query Your Notes