File Format
NOTE: This article is intended for programmers who want to write programs that interface with budget files. You don't need this if you just want to use the Buckets app.
Buckets budget files are SQLite databases. The schema is mostly self-explanatory, but this document points out some things that aren’t.
WARNING: The schema is not final and may change without warning in future versions of Buckets.
If you still have questions after reading this, feel free to chat with me.
Concurrency¶
It’s not a good idea to do data-manipulating operations (INSERT
, UPDATE
, DELETE
, etc…) on the same budget file from more than one process at a time. So before you manipulate data outside of Buckets, close the budget file inside Buckets.
Amounts¶
Amounts are stored as integers, not floats. For USD (and most other currencies), this means that amounts are integer cents.
In this example, account 1 has a balance of $6,500:
sqlite> SELECT balance FROM account WHERE id=1;
650000
sqlite> SELECT printf("%.2f", balance/100.0) AS balance FROM account WHERE id=1;
6500.00
Inserting transactions¶
Insert bank/account transactions into the account_transaction
table. Be aware of the following:
- The
fi_id
column represents an account-unique ID (typically assigned by the bank) for a transaction. general_cat
should be one of the strings""
,"income"
, or"transfer"
.
Account and bucket balances are automatically updated (by SQLite triggers) when transactions are inserted/updated/deleted.
Custom tables, views and triggers¶
You are welcome to add custom tables, views and triggers to your budget file at your own risk. Please follow these conventions:
- Prefix the name of all custom things with
c_
(e.g.CREATE TABLE c_mytable ...
) - When merging budgets, custom tables are combined additively. So you may need to manually delete records in both budgets.