Search...Search plugins and themes...
⌘K
Sign in
  • Get started
  • Download
  • Pricing
  • Enterprise
  • Account
  • Obsidian
  • Overview
  • Sync
  • Publish
  • Canvas
  • Mobile
  • Web Clipper
  • CLI
  • Learn
  • Help
  • Developers
  • Changelog
  • About
  • Roadmap
  • Blog
  • Resources
  • System status
  • License overview
  • Terms of service
  • Privacy policy
  • Security
  • Community
  • Plugins
  • Themes
  • Discord
  • Forum / 中文论坛
  • Merch store
  • Brand guidelines
Follow us
DiscordTwitterBlueskyThreadsMastodonYouTubeGitHub
© 2026 Obsidian

CFR Js

cferrugemcferrugem26 downloads

Fast vault queries (TABLE/LIST/TASK + JavaScript API) in the style of Dataview, built for performance.

Add to Obsidian
  • Overview
  • Scorecard
  • Updates2

An Obsidian plugin inspired by Dataview, written from scratch with a focus on performance: TABLE / LIST / TASK queries, a JavaScript API, inline queries, inline fields, task toggling from views, and CSV sources.

cfrjs is an independent project. It is not affiliated with or endorsed by Dataview or its author.

Manual installation

npm install
npm run build        # type-checks and produces main.js

Copy main.js, manifest.json and styles.css to <your vault>/.obsidian/plugins/cfrjs/ and enable the plugin under Settings → Community plugins.

Development: npm run dev (rebuild on change), npm test, npm run bench.

Queries (```cfr)

```cfr
TABLE status, priority AS "Priority", file.mtime
FROM #project AND "Work"
WHERE priority >= 3 AND !contains(file.tags, "#archived")
SORT file.mtime DESC
LIMIT 20
```
Command Example
TABLE [WITHOUT ID] expr [AS name], ... TABLE status, length(file.tasks) AS Tasks
LIST [WITHOUT ID] [expr] LIST file.mtime
TASK TASK WHERE !completed
FROM source #tag, "folder", [[note]] (incoming links), outgoing([[note]]), csv("data.csv"), combined with and, or, -/! and parentheses
WHERE expr WHERE due <= date(today) + dur(7 days)
SORT expr [ASC|DESC], ... SORT priority DESC, file.name
GROUP BY expr [AS name] GROUP BY status (produces key and rows)
FLATTEN expr [AS name] FLATTEN file.tags
LIMIT n LIMIT 10

Expressions: + - * / %, = != < <= > >=, and/or/!, field.sub, list[0], row["field with spaces"], this (current page), links [[Note]].field, lists [1, 2], objects { a: 1 }, lambdas (x) => x.y, dates date(today), date(2024-05-01) and durations dur(1 week).

Page fields: frontmatter, inline fields (field:: value, [field:: value], (field:: value)) and file.name, file.path, file.folder, file.link, file.tags, file.etags, file.aliases, file.outlinks, file.inlinks, file.tasks, file.lists, file.ctime, file.cday, file.mtime, file.mday, file.size, file.day, file.frontmatter, file.starred. Field names are normalized (Due Date:: is also available as due-date).

Tasks/lists: text, status, checked, completed, fullyCompleted, line, path, section, link, tags, outlinks, children, parent, due, created, start, scheduled, completion (including emoji shorthand 📅 ✅ ⏳ 🛫 ➕) and the item's inline fields. Page fields are also visible in TASK queries.

Functions: date, dur, number, string, link, embed, elink, typeof, object, list, meta, display, round, trunc, floor, ceil, abs, min, max, minby, maxby, sum, product, average, currencyformat, length, contains, icontains, econtains, containsword, extract, reverse, sort, flat, slice, unique, nonnull, firstvalue, filter, map, reduce, any, all, none, join, default, ldefault, choice, get, lower, upper, trim, split, replace, regextest, regexmatch, regexreplace, startswith, endswith, padleft, padright, substring, truncate, striptime, dateformat, durationformat, localtime.

JavaScript (```cfrjs)

```cfrjs
const projects = cfr.pages("#project").where(p => p.status === "active");
cfr.header(3, `${projects.length} active projects`);
cfr.table(
    ["Project", "Priority", "Open"],
    projects.sort(p => p.priority, "desc").map(p => [p.file.link, p.priority, p.file.tasks.where(t => !t.completed).length])
);
```

dv also works as an alias of cfr, to ease migration.

API Description
cfr.pages(source?), cfr.pagePaths(source?), cfr.page(path), cfr.current() Index data
cfr.query(dql) / cfr.tryQuery(dql) Run DQL and return data
cfr.execute(dql) Render a DQL query
cfr.table(headers, rows), cfr.list(items), cfr.taskList(tasks, groupByFile?) Rendering
cfr.paragraph, cfr.header, cfr.span, cfr.el(tag, content, {cls, attr}) Elements
cfr.io.csv(path), cfr.io.load(path), cfr.view(path, input) Files
cfr.date, cfr.duration, cfr.fileLink, cfr.sectionLink, cfr.blockLink, cfr.compare, cfr.equal, cfr.array, cfr.func.*, cfr.evaluate(expr) Utilities

DataArray is a real Array subclass with where, sort(key, "asc"|"desc"), groupBy, distinct, limit, first, last, to, into, pluck("file.name"), expand("children"), sum, avg, min, max, none, array() and chained access to the most common fields (pages.file.name, page.file.tasks.text).

Other plugins can use window.CfrJsAPI (pages, page, query, stats).

Inline

  • `= this.status` — expression evaluated in Reading view and Live Preview.
  • `$= cfr.pages("#project").length` — inline JavaScript.

Commands

  • Refresh all views
  • Rebuild index (discard cache)
  • Show index statistics — indexing times, cache hits, workers and how many refreshes were skipped.

Why it is faster

Startup and indexing

  • The persistent cache is loaded with a single IndexedDB read, and changes are written back in batched transactions.
  • Frontmatter, tags and links come straight from Obsidian's metadata cache, so only files without a valid cache entry are read from disk.
  • Inline fields and lists are parsed by up to 4 Web Workers (configurable), in batches of 64 files, walking each file's lines once.
  • Index updates are coalesced and emitted in batches instead of one notification per file.

Queries

  • Each page's queryable row is built once per version and shared by every query; file.* values are computed lazily and memoized.
  • Tags, folders, incoming links and unresolved links have incremental indexes, so FROM never scans the whole vault.
  • Expressions are compiled to JavaScript closures once and reused on every refresh.
  • Sorting uses a typed comparator with fast paths; SORT ... LIMIT selects the top results with a heap instead of sorting everything; GROUP BY uses hash maps; FLATTEN creates shallow overlays instead of copies.
  • DataArray is a real Array subclass, so indexing and iteration have native cost.

Refreshing and rendering

  • Every view records what it depends on (tags, folders, files, links, CSV files); a change only refreshes the views it affects, and off-screen views wait until they become visible.
  • Results are rendered directly to the DOM; the Markdown renderer is used only for text that actually contains Markdown.
  • Unchanged rows are reused between refreshes, and large results are rendered in chunks during idle time.
  • In Live Preview, only the visible part of the document is scanned, and inline results are recomputed only when their code or dependencies change.
  • No runtime dependencies: the whole plugin, worker included, is about 115 KB minified.

Differences from Dataview

  • Block keywords: cfr and cfrjs (enable Accept dataview/dataviewjs blocks to migrate without editing notes, with Dataview disabled).
  • Dates are CDate (not Luxon). Common methods exist: year, month, day, weekday, toFormat(), toISODate(), toMillis(), plus(), minus(), startOf(), endOf(). dv.luxon is not available.
  • DataArray.sort(key) returns a new array; sort((a, b) => ...) with two arguments keeps the native (in-place) behavior.
  • Page objects are shared between queries to avoid copies: do not mutate them in scripts (create new objects instead).
  • Not implemented in this version: CALENDAR, pretty-rendering of inline fields in Live Preview, recursive subtask completion.

Security

cfrjs code blocks and inline $= expressions run JavaScript written in your notes with the same permissions as Obsidian itself (they can read and modify files in your vault). Only open vaults and notes you trust. You can turn JavaScript off under Settings → CFR Js → Enable JavaScript and Enable inline JavaScript; cfr queries and inline = expressions never execute arbitrary code.

Mobile

The plugin does not use Node.js or Electron APIs and falls back to the main thread when Web Workers are unavailable, so it is marked as mobile compatible. Mobile has not been extensively tested yet; please report issues.

Credits and license

cfrjs is released under the MIT License.

Its query language, API names and field names follow Dataview's conventions so existing queries keep working. The inline field parsing rules in src/index/parse-content.ts are adapted from Dataview (Copyright (c) 2021 Michael Brenan, MIT License); the full notice is included in THIRD_PARTY_NOTICES.md.

Releasing

  1. npm version patch (or minor/major) — updates package.json, manifest.json and versions.json and creates a git tag.
  2. git push --follow-tags — the GitHub Action builds, tests and creates a draft release with main.js, manifest.json and styles.css.
  3. Review and publish the draft release on GitHub.

Benchmarks

npm run bench measures the core in Node (in-memory index with 10,000 notes). To test inside Obsidian:

node scripts/gen-vault.mjs ./test-vault 10000

Open test-vault as a vault, install the plugin and use Show index statistics and the Show timings option.

HealthExcellent
ReviewPassed
About
Execute TABLE, LIST, and TASK queries with a high-performance JavaScript engine, inline queries and fields, CSV sources, and a JavaScript API for custom query logic. Toggle tasks directly from query views and use expressions, dates/durations, grouping, flattening and combined sources to build flexible data-driven views.
TablesTasksData
Details
Current version
0.1.1
Last updated
Last week
Created
Last week
Updates
2 releases
Downloads
26
Compatible with
Obsidian 1.4.0+
Platforms
Desktop, Mobile
License
MIT
Report bugRequest featureReport plugin
Sponsor
Buy Me a Coffee
Author
cferrugemcferrugemcferrugem
GitHubcferrugem
  1. Community
  2. Plugins
  3. Tables
  4. CFR Js

Related plugins

Datacore

An even faster reactive query engine for the data obsessed.

dotpm

Full-featured task and project management: stunning Gantt charts, Kanban boards, Table views, customizable fields, due date notifications, dependancies

TileLineBase

Build multidimensional tables (Tables, Kanban, Gallery, Slides) directly from plain text. Zero Frontmatter. Zero code.

Text Block Timer

Add a timer to text block to track task time consumption.

DataLoom

Weave together data from diverse sources into different views. Inspired by Excel Spreadsheets and Notion.so.

TaskNotes

Note-based task management with calendar, pomodoro and time-tracking integration.

Advanced Tables

Improved table navigation, formatting, and manipulation.

Tasks

Track tasks across your vault. Supports due dates, recurring tasks, done dates, sub-set of checklist items, and filtering. Maintained by Clare Macrae and Ilyas Landikov, created by Martin Schenck.

Day Planner

Turn tasks from daily notes, the Tasks plugin, and calendars into time blocks on an editable timeline, with built-in time tracker.

Tracker

Track occurrences and numbers in your notes.