Knowledge Concepts
Full-text search is the ability to search every word in every document in a collection — not just titles or tags — returning results ranked by relevance to the query. It's the difference between a filing cabinet and a search engine.
Full-text search is the ability to search across the actual content of every document in a collection — every word in every file — rather than just searching metadata like titles, filenames, or tags. A full-text search for "cognitive load" finds every note, article, or document where those words appear, even if they're never in the title or tags. Full-text search is what makes a large collection of documents findable without perfect upfront organization.
Full-text search is the difference between a filing cabinet (you must know exactly where you put something) and a search engine (you can describe what you're looking for).
Early database systems searched only structured fields (indexed columns) — not the unstructured text content of documents. Searching document content required reading every document sequentially, which was too slow for large collections.
Two key developments made full-text search practical:
Inverted indexes (1960s-1970s): An inverted index maps each word to the list of documents containing it — essentially reversing the document→word relationship. Instead of scanning every document for the word "cognitive," a search queries the inverted index and retrieves the pre-computed list of documents containing "cognitive" in milliseconds. This is the foundational data structure behind all modern search.
TF-IDF and relevance ranking (1970s-1980s): Term Frequency-Inverse Document Frequency (TF-IDF) is a statistical measure for ranking documents by relevance to a query. TF-IDF balances how often a term appears in a document (term frequency) against how commonly it appears across all documents (inverse document frequency). A term that appears 10 times in a document but appears in every document ranks lower than one that appears 5 times in a rare document. This relevance ranking made search results actually useful.
Commercial and open-source implementations: Lucene (1999) — the open-source Java search library that powers Elasticsearch, Solr, and many others. Elasticsearch (2010) — a distributed search system built on Lucene, now ubiquitous. Algolia (2012) — hosted search-as-a-service for developers. All modern note-taking apps with good search use variants of these approaches.
Step 1 — Indexing: When you save a document, the search system processes its text:
Step 2 — Querying: When you search "cognitive load":
Step 3 — Features: Modern full-text search systems add:
A developer has 400 notes in Obsidian — meeting notes, debugging notes, ADRs, code snippets, and article summaries. They remember reading something about "memory leak" debugging in a Node.js context, but they can't remember which note or when.
Without full-text search: Navigate folders manually. Browse recent notes. Nothing. Check a few folder categories. Still nothing. Spend 15 minutes not finding it.
With full-text search (Obsidian's built-in or any note app with FTS): Search "memory leak Node.js" — results appear in 0.3 seconds. Three notes match. The first is the debugging session from 6 months ago with the exact solution they needed.
Total time: 5 seconds. The note was buried three folders deep under a name they'd never have remembered — but the full-text index found it by content.
| Search type | How it works | When it finds things |
|---|---|---|
| Filename/title search | Matches against document titles only | Only if you remember the exact title |
| Metadata/tag search | Matches against tags, categories, custom fields | Only if you applied the right tags |
| Full-text search | Matches against every word in every document | Whenever the words appear, regardless of tags |
| Semantic/vector search | Matches by meaning, not exact words | When you describe the concept without using the exact words |
Full-text search is the critical layer that makes collections findable even with imperfect tagging. Not every note has perfect metadata; every note has its actual content.
Most modern note-taking tools include full-text search. Quality varies significantly:
| Tool | Full-text search quality | Notes |
|---|---|---|
| Obsidian | Excellent | Fast, regex support, field-specific search |
| Notion | Good | Works well; can be slower on large databases |
| Evernote | Good | Has long had strong FTS; one of its original advantages |
| Apple Notes | Good | Fast native search |
| Roam Research | Good | Real-time FTS |
| Bear | Good | FTS with tag and content filtering |
| Standard Notes | Basic | Title + content search |
What to look for in a note app's search:
Full-text search changes the optimal information architecture for personal knowledge systems:
Without full-text search: You must organize meticulously — every note needs detailed tags and a descriptive filename, because you can only find things through navigation and metadata. The cost of poor organization is total invisibility.
With full-text search: Imperfect metadata is less costly — the content itself is searchable. You can write naturally in notes without exhaustive tagging, because the search will find what you wrote even without the right tag.
The implication: Full-text search doesn't eliminate the value of good organization — categories and tags still help you browse and filter. But it changes the stakes: a note with no tags but good content is findable; a note with perfect tags but no substantive content adds nothing.
"Full-text search replaces the need for organization." Full-text search reduces the penalty for poor organization, but doesn't eliminate the value of good organization. Tags and categories help you browse and filter when you don't know the exact words to search for. Full-text search handles the "I know it exists" case; organization handles the "I'm exploring a topic" case.
"Full-text search only works if you use the exact word." Modern full-text search includes stemming (finding "search" when you type "searching"), synonyms (configurable), and fuzzy matching (finding "cogntive" when you type "cognitive"). Exact word matching is the basic implementation; good implementations are much more flexible.
"Full-text search is only for developers." Full-text search is in Gmail, Notion, Obsidian, Evernote, Slack, and virtually every information tool you use daily. Understanding how it works helps you use it better — knowing that searches for "memory leak" and "memory leaks" may return different results, or that phrase searches use quotes, gives you more precise control.
Inverted index: The data structure that makes full-text search fast — maps words to document lists rather than vice versa.
Vector search / semantic search: The newer approach that finds documents by meaning rather than exact words. Complementary to full-text search for different query types.
Metadata: Structured document descriptions that enable field-specific searches and filtering — complements full-text search.
How is full-text search different from grep?
grep searches file contents using regex pattern matching — scanning files sequentially. It's powerful but slow on large collections because it reads every file. Full-text search uses a pre-built inverted index and is orders of magnitude faster on large collections. grep is a Unix tool; full-text search is a system with indexing infrastructure.
Does full-text search index attachments (PDFs, images)? Depends on the tool. Some search engines extract text from PDFs and index it. Images (without OCR) can't be full-text searched because there's no text to index. Tools like DevonThink include OCR to extract text from images and PDFs for indexing. Most note-taking apps index only text content.
How do I use full-text search more effectively?
Key techniques: use phrase search (quotes: "memory leak") for exact phrases; use Boolean operators (AND, OR, NOT) where supported; use field-specific search (title:cognitive to search titles only); search for distinctive phrases rather than common words; understand stemming (searching "search" also finds "searches," "searching").
Full-text search is the technology that makes large information collections manageable. Without it, findability depends entirely on perfect upfront organization — which is impossible to maintain at scale. With full-text search, imperfect notes in imperfect folders are still findable when you search for their content. For developers and knowledge workers who accumulate hundreds or thousands of notes, debugging logs, ADRs, and article summaries: the presence or absence of good full-text search in their tools is one of the highest-leverage decisions they make.
More WebSnips articles that pair well with this topic.
Intermediate packets are small, discrete, reusable units of work — an outline, a research summary, a slide deck draft, a set of notes — that can be saved and reused across future projects rather than starting from scratch each time.
A bullet journal (BuJo) is an analog organization system created by Ryder Carroll using a dot-grid notebook: rapid logging with bullet symbols, collections (logs, trackers, lists), and migration (intentional task review). It combines daily planner, to-do list, diary, and habit tracker in one notebook.
A KWL chart is a three-column learning organizer: K (What I Know), W (What I Want to Know), L (What I Learned). It activates prior knowledge before learning, sets specific goals, and documents what you actually learned — making the learning process active and self-directed.