A primitive file-based set supporting CRUD operations.

This small tool serves as a minimal file-based persistent unique key database supporting CRUD operations on the key. Assuming infrequent access by a single application, there's no file locking.

Many applications implement this on their own (and usually within a few lines with grep); by delegating to this tool, one gets a robust yet small implementation, consistency in usage and storage locations, and the possibility to easily upgrade to the more powerful yet similar nanoDB and miniDB APIs.

Each database "table" is represented as an individual file (put by default under ~/.local/share/[NAMESPACE/]TABLE; the location can be customized via command-line arguments or $XDG_DATA_HOME). Each record is a line consisting of a newline-escaped KEY (so newlines and any other special character can be used). Keys can be queried separately for existence, or the whole database can be gotten, either as the lines (newlines have to be unescaped by the client) or as a Bash associative array definition that can be eval'ed into existence and then tested for keys.

• Bash, grep, GNU sed
• automated testing is done with bats - Bash Automated Testing System (https://github.com/bats-core/bats-core)

• The ./bin subdirectory is supposed to be added to PATH.

• nanoDB offers association of a single value for every key, with a very similar API.
• miniDB offers multiple columns and transaction / file locking support behind a very similar API.