What features would you want for a `q` Querystring parser? (e.g. full-text-search, or more?)
Pacharapol Withayasakpunt
Posted on December 3, 2020
It might also govern which database you will choose... just MongoDB with full-text-search, or do you need Elasticsearch or something else? Will the data be duplicated?
Not sure about PostGres, MySQL, MariaDB, but SQLite does have an awesome full-text-search engine (e.g. FTS5).
CREATE TABLE IF NOT EXISTS files (
id TEXT PRIMARY KEY,
lastRead FLOAT, -- +new Date() number; nullable
rating FLOAT DEFAULT 0
);
CREATE INDEX IF NOT EXISTS files_lastRead ON files(lastRead);
CREATE INDEX IF NOT EXISTS files_rating ON files(rating);
CREATE VIRTUAL TABLE IF NOT EXISTS q USING fts5(
fileId, -- REFERENCES files(id)
author, -- ', ' joined Set
title, -- ', ' joined Set
tag, -- ', ' joined Set
frontmatter, -- other searchable frontmatter
content -- cleaned markdown-to-plaintext
);
But, is only full-text-search really enough? Some of other features I would be dying for includes
- Search by fields
- Number fields
- Datetime fields
- Boolean operators / brackets
- Wildcards
- Fuzzy search
- Language-specific normalizations
patarapolw
/
qsearch
Search a database with a string. Designed for end-users.
Personally, I don't care about field boosting much, but it is featured in some search engines.
💖 💪 🙅 🚩
Pacharapol Withayasakpunt
Posted on December 3, 2020
Join Our Newsletter. No Spam, Only the good stuff.
Sign up to receive the latest update from our blog.
Related
discuss What features would you want for a `q` Querystring parser? (e.g. full-text-search, or more?)
December 3, 2020