How to use the Obsidian Dataview plugin

Obsidian Dataview turns an entire Obsidian vault into a queryable database, letting notes behave less like a static archive and more like structured knowledge that can be searched, filtered, and reshaped on demand. The core promise is practical: instead of building a “second brain” that merely mirrors the first brain’s weaknesses at retrieval and storage, Dataview supports a bottom-up note-taking workflow where raw notes accumulate—then become usable through metadata and live queries.

Dataview is an Obsidian community plugin that converts notes into database records. It can feel daunting because metadata and queries add complexity, but the workflow can stay simple: annotate only what matters, then query only what you need. Installation is straightforward—open Obsidian Settings, disable safe mode if necessary, browse Community Plugins, install Dataview, and enable it.

Using Dataview has two distinct steps. First comes annotation: each note can be labeled with fields that Dataview will index. Those fields can be written directly in the note body using “double colons,” or hidden in YAML front matter using three dashes at the top and bottom. Either way, the parameters are treated the same; the YAML approach just keeps metadata out of the visible preview. In the example, a “person” note for Kieran includes fields like type, location, date_met, tags such as D&D, and a project assignment like Project X.

Second comes querying. Dataview queries live inside code blocks that begin and end with three backticks, and the block is tagged as either dataview (DQL) or dataview JS. Queries start by choosing a view type—lists, tables, tasks, or calendars—then selecting which annotated fields to display. Results can be limited by adding a FROM clause (such as a folder like “Meetings” or a tag filter like #meeting) and narrowed further with WHERE criteria based on the metadata fields. SORT can order results, commonly with file.name ASC for alphabetical ordering.

The transcript walks through how small metadata changes update multiple views automatically. A Project X page can generate a list or table of people assigned to that project using WHERE type is person AND project = "Project X". When a meeting note is created with fields like meeting, project, attendees, and summary, a Dataview table on the Project X page can pull meeting date and summary—either by filtering with type = meeting or by restricting the query to the Meetings folder. A glossary example shows how tags can drive retrieval: a performance testing glossary table lists terms and definitions only when tags include “glossary” and optionally when tags contain “performance.”

Other examples show Dataview’s breadth: a “books I read in 2022” page filters Readwise-imported notes by date (date >= 2022-01-01) and displays rating, title, and date; a tasks view pulls only incomplete tasks tied to Project X; and a calendar view uses file.mtime to show recently modified notes, again filtered by tags like #books. The takeaway is that Dataview helps prevent hoarding by turning scattered notes into a living database of what’s actually been learned—without requiring notes to be perfectly categorized at creation time.