Obsidian Dataview

To develop plugins that interact with Dataview, install the obsidian-dataview package as a development dependency. You can verify the installed version using npm list obsidian-dataview. If you need to ensure you are using a specific version (e.g., 0.5.64), you can install it explicitly.

Note: Git must be installed on your local system before proceeding.

npm install -D obsidian-dataview

# To verify version
npm list obsidian-dataview

# To install a specific version
npm install obsidian-dataview@0.5.64

The default type for any field is Text.

For inline fields, a line break signifies the end of the value.

To create multiline text, you must use YAML Frontmatter with the pipe (|) operator.

---
poem: |
  Because I could not stop for Death,
  He kindly stopped for me;
---

When using GROUP BY, a standard LIST query will only show the group keys. To see the files within those groups, you must specify rows.file.link as the additional information.

LIST rows.file.link
GROUP BY type

The FLATTEN command takes an array in a row and yields one result row for every entry in that array. This is useful for operating on nested lists like file.lists or file.tasks without complex map() or filter() functions.

Syntax:

FLATTEN field
FLATTEN (computed_field) AS name

Example: To list every author in literature notes as an individual row:

TABLE authors FROM #LiteratureNote
FLATTEN authors

Most Dataview functions support "function vectorization." This means if you apply a function to a list of values, the function is applied to each element individually, and the result is a new list of the same length. This works for both single values and lists.

lower("YES") = "yes"
lower(["YES", "NO"]) = ["yes", "no"]

replace("yes", "e", "a") = "yas"
replace(["yes", "ree"], "e", "a") = ["yas", "raa"]

If you have a field name that is identical to a Dataview keyword (such as from or where), you cannot access it directly. Instead, use the special row object to index into the field name.

row.from /* Same as "from" */
row.where /* Same as "where" */

Use the AS <header> syntax to rename columns. If a header contains spaces, wrap it in double quotes (e.g., AS "File Tags"). This is particularly useful when using calculations as column values.

TABLE default(finished, date(today)) - started AS "Played for",
file.folder AS Path,
file.etags AS "File Tags"
FROM #games

How you access task-specific implicit fields depends on the query type:

In TASK queries

Tasks are the top-level objects, so you can access fields directly without a prefix.

TASK
WHERE !fullyCompleted

In other query types (LIST, TABLE, etc.)

You must access the fields through the file.tasks or file.lists collection using a list function.

LIST
WHERE any(file.tasks, (t) => !t.fullyCompleted)

DataviewJS provides a high-powered JavaScript API for full access to the Dataview index and custom rendering. It is more powerful than DQL but requires JavaScript knowledge.

Security Note: JavaScript queries run at the same privilege level as Obsidian plugins and can potentially modify files or make network calls. Use them only with trusted code.

Example: List all uncompleted tasks:

dv.taskList(dv.pages().file.tasks.where(t => !t.completed));

Use the `

TABLE file.ctime, status
FROM "projects/brainstorming"