A lightweight, SQLite-compatible version of the JMdict Japanese-English dictionary.
Automatic daily releases @ 5AM UTC.
Download latest: https://github.com/seanmcbroom/JMdictSQLite/releases/download/latest/jmdict.sqlite
This project transforms the JMdict XML dataset into a normalized, developer-friendly SQLite format for use in apps, browser extensions, or any application needing a fast, local Japanese-English dictionary.
- Designed for easy integration into apps and tools
- Compact schema with JSON-encoded kanji, kana, and sense data
- Ideal for full-text search, filtering, and offline usage
PK = Primary Key
FK = Foreign Key
? = possibly undefined
entries.ent_seq → senses.ent_seq (one-to-many)
┌────────────────────────────────────┐ ┌───────────────────────────────────────────────────┐
│ entries │◄───────────│ senses │
├────────────────────────────────────┤ ├───────────────────────────────────────────────────┤
│ ent_seq INTEGER PK │◄────┐ │ id INTEGER PK AUTOINC │
│ kanji? TEXT │ └────► │ ent_seq INTEGER FK → entries.ent_seq │
│ – JSON-encoded │ | note? TEXT |
│ kana TEXT │ │ glosses TEXT (JSON-encoded) │
│ – JSON-encoded │ │ pos TEXT (JSON-encoded) │
└────────────────────────────────────┘ | verb_data? TEXT (JSON-encoded) |
| field? TEXT (JSON-encoded) |
| tags? TEXT (JSON-encoded) |
└───────────────────────────────────────────────────┘
export interface EntryDb {
ent_seq: number; // PK
kanji?: string; // JSON.stringify(Written[])
kana: string; // JSON.stringify(Written[])
}
export interface SenseDb {
id: number; // PK
ent_seq: number; // FK
note?: string;
glosses: string; // JSON.stringify(string[])
pos: string; // JSON.stringify(string[])
verb_data?: string; // JSON.stringify(VerbData)
fields?: string; // JSON.stringify(string[])
tags?: string; // JSON.stringify(string[])
}
interface Written {
written: string
tags?: []
}
interface VerbData {
verb_group: string;
transivity?: string;
}This SQL query retrieves all entries that contain at least one of the specified tags (ichi1 or nf01) in any of their associated kanji objects.
It works by iterating through the kanji array of each entry using json_each, and then checking whether the tags array within each kanji object contains any of the target tags.
To perform an inclusive (AND) search, where all specified tags must be present, change >= 1 to = 2 (or however many tags you're searching for).
SELECT *
FROM entries
WHERE EXISTS (
SELECT 1
FROM json_each(entries.kanji) AS kanji_obj
WHERE (
SELECT COUNT(DISTINCT tag.value)
FROM json_each(json_extract(kanji_obj.value, '$.tags')) AS tag
WHERE tag.value IN ('ichi1', 'nf01')
) >= 1
);Requires: node 22.11.00+, npm 10.9.0+
git clone https://github.com/seanmcbroom/JMdictSQLitecd JMdictSQLitenpm installnpm run create-releaseThe resulting SQLite file will be generated at: /data/jmdict.sqlite
You are free to use or distribute the file in accordance with the terms of the GPLv2 license.
This project contains components originally created by the Electronic Dictionary Research and Development Group (EDRDG). These components are derived from data licensed under the GNU General Public License Version 2 (GPLv2).
You are free to use, modify, and distribute this software under the terms of the GPLv2. A copy of the full license is included in the LICENSE file.