Obsidian Basics: Search & Starred Search

Obsidian’s search just got much more powerful with version 0.8.0, adding regex support so users can match patterns instead of only plain words. The practical payoff is faster, more precise retrieval—whether filtering by file name or path, enforcing case rules, or hunting structured data like dates, amounts, phone numbers, emails, URLs, and even sleep logs.

Search starts from the magnifying glass in the top-left, where three controls shape results before any advanced syntax kicks in. “Collapse results” switches between showing only matching node names and showing the lines containing matches. “Match case” toggles case-sensitive matching, while “Explain search term” reveals what the query is doing in plain language—useful when debugging complex regex patterns.

Basic searching remains straightforward: typing a word returns notes containing that word anywhere in the file. Multiple words default to an “anywhere” match across the note, but quoting a phrase forces consecutive terms (e.g., wrapping BCP together in quotes). For more complex logic, boolean operators let users combine or exclude matches. Using “or” returns notes containing either term. A leading “-” negates results that contain a term, and parentheses group conditions so “A and B” can be combined with exclusions like “not C.”

Four special operators extend search beyond the text body. “file:” searches the file name, “path:” searches the absolute file path from the vault root, “match case:” enforces case sensitivity for that query, and “ignore case:” flips it to case-insensitive matching. These operators can be layered with the rest of the search syntax.

The biggest change is regex: patterns are wrapped in slashes, and character classes and metacharacters require careful escaping. The transcript lists common classes such as “.” (any character except newline), “\d” (digits), “\D” (non-digits), “\w” (word characters), “\W” (non-word characters), “\s” (whitespace), and “\S” (non-whitespace). Square brackets define sets (e.g., [abc]) and negated sets use a caret (e.g., [^abc]). Quantifiers control repetition: “*” means 0 or more, “+” means 1 or more, “?” means 0 or 1, and curly braces specify exact counts or ranges like {1,3}.

Examples show how regex becomes a data tool. Escaping “$” allows matching dollar amounts like “$25” or “$125” using “\d{1,3}.” Phone numbers are handled by escaping “+” and combining country codes with digit ranges (e.g., “\+91 \d{10}” and variants for other formats). Email matching uses letter ranges and repetition (e.g., [a-z]+@[a-z]+\.[a-z]+), while URLs start with “http” and optionally “s,” then use word-character runs and escaped dots to capture domains. Even personal tracking is addressed: sleep entries are matched with patterns like “\d{1,2}H \d{1,2}M,” with optional markers to include days where hours or minutes are missing (using “?” to make parts optional).

Finally, the workflow improvement isn’t just search syntax: a “Star” plugin can save the current search so the same regex can be rerun later with a single click—turning one-off queries into reusable filters for recurring tasks like finding all sleep notes.