sqlite-export-for-ynab 2.0.0

SQLite Export for YNAB - Export YNAB Data to SQLite

sqlite-export-for-ynab

SQLite Export for YNAB - Export YNAB Budget Data to SQLite

What This Does

Export all your YNAB plans to a local SQLite DB. Then you can query your data with any tools compatible with SQLite.

Installation

$ pip install sqlite-export-for-ynab

Usage

CLI

Provision a YNAB Personal Access Token and save it as an environment variable.

$ export YNAB_PERSONAL_ACCESS_TOKEN="..."

Run the tool from the terminal to download your plans:

$ sqlite-export-for-ynab

Running it again will pull only data that changed since the last pull (this is done with Delta Requests). If you want to wipe the DB and pull all data again use the --full-refresh flag.

You can specify the DB path with the following options

1. The --db flag.
2. The XDG_DATA_HOME variable (see the XDG Base Directory Specification). In that case the DB is saved in "${XDG_DATA_HOME}"/sqlite-export-for-ynab/db.sqlite.
3. If neither is set, the DB is saved in ~/.local/share/sqlite-export-for-ynab/db.sqlite.

Library

The library exposes the package sqlite_export_for_ynab and two functions - default_db_path and sync. You can use them as follows:

import asyncio
import os

from sqlite_export_for_ynab import default_db_path
from sqlite_export_for_ynab import sync

db = default_db_path()
token = os.environ["YNAB_PERSONAL_ACCESS_TOKEN"]
full_refresh = False

asyncio.run(sync(token, db, full_refresh))

Relations

The relations are defined in create-relations.sql. They are 1:1 with YNAB's OpenAPI Spec (ex: transactions, accounts, etc) with some additions:

1. Some objects are pulled out into their own tables so they can be more cleanly modeled in SQLite (ex: subtransactions, loan account periodic values).
2. Foreign keys are added as needed (ex: plan ID, transaction ID) so data across plans remains separate.
3. Two new views called flat_transactions and scheduled_flat_transactions. These allow you to query split and non-split transactions easily, without needing to also query subtransactions and scheduled_subtransactions respectively. They also include fields to improve quality of life (ex: amount_major to convert from YNAB's milliunits to major units i.e. dollars) and filter out deleted transactions/subtransactions.

Querying

You can issue queries with typical SQLite tools. sqlite-export-for-ynab deliberately does not implement a SQL REPL.

Sample Queries

To get the top 5 payees by spending per plan, you could do:

WITH
ranked_payees AS (
    SELECT
        pl."name" AS plan_name
        , t.payee_name AS payee
        , SUM(t.amount_major) AS net_spent
        , ROW_NUMBER() OVER (
            PARTITION BY
                pl.id
            ORDER BY
                SUM(t.amount) ASC
        ) AS rnk
    FROM
        flat_transactions AS t
    INNER JOIN plans AS pl
        ON t.plan_id = pl.id
    WHERE
        t.payee_name != 'Starting Balance'
        AND t.transfer_account_id IS NULL
    GROUP BY
        pl.id
        , t.payee_id
)

SELECT
    plan_name
    , payee
    , net_spent
FROM
    ranked_payees
WHERE
    rnk <= 5
ORDER BY
    plan_name ASC
    , net_spent DESC
;

To get duplicate payees, or payees with no transactions:

SELECT DISTINCT
    pl."name" AS "plan"
    , dupes."name" AS payee
FROM (
    SELECT DISTINCT
        p.plan_id
        , p."name"
    FROM payees AS p
    LEFT JOIN flat_transactions AS ft
        ON
            p.plan_id = ft.plan_id
            AND p.id = ft.payee_id
    LEFT JOIN scheduled_flat_transactions AS sft
        ON
            p.plan_id = sft.plan_id
            AND p.id = sft.payee_id
    WHERE
        TRUE
        AND ft.payee_id IS NULL
        AND sft.payee_id IS NULL
        AND p.transfer_account_id IS NULL
        AND p."name" != 'Reconciliation Balance Adjustment'
        AND p."name" != 'Manual Balance Adjustment'
        AND NOT p.deleted

    UNION ALL

    SELECT
        plan_id
        , "name"
    FROM payees
    WHERE NOT deleted
    GROUP BY plan_id, "name"
    HAVING COUNT(*) > 1

) AS dupes
INNER JOIN plans AS pl
    ON dupes.plan_id = pl.id
ORDER BY "plan", payee
;

To count the spend for a category (ex: "Apps") between this month and the next 11 months (inclusive):

SELECT
    plan_id
    , SUM(amount_major) AS amount_major
FROM (
    SELECT
        plan_id
        , amount_major
    FROM flat_transactions
    WHERE
        category_name = 'Apps'
        AND SUBSTR(`date`, 1, 7) = SUBSTR(DATE(), 1, 7)
    UNION ALL
    SELECT
        plan_id
        , amount_major * (
            CASE
                WHEN frequency = 'monthly' THEN 11
                ELSE 1 -- assumes yearly
            END
        ) AS amount_major
    FROM scheduled_flat_transactions
    WHERE
        category_name = 'Apps'
        AND SUBSTR(date_next, 1, 7) < SUBSTR(DATE('now', '+1 year'), 1, 7)
)
;