clj-yfinance.duckdb — com.github.clojure-finance/clj-yfinance 0.1.6

Optional DuckDB integration for clj-yfinance datasets.

Loads datasets fetched via clj-yfinance into an embedded DuckDB instance and lets you query them with SQL.

REQUIRES: Add to your deps.edn aliases: {:duckdb {:extra-deps {com.techascent/tmducken {:mvn/version "0.10.1-01"} techascent/tech.ml.dataset {:mvn/version "7.032"}}}}

DuckDB also requires a native shared library (libduckdb). On Linux you can install it via your package manager, or set the DUCKDB_HOME environment variable to the directory containing the library before starting your REPL.

Then start your REPL with: clojure -M:duckdb:nrepl

USAGE: (require '[clj-yfinance.duckdb :as yf-db])

;; Open an in-memory database and load some data (def db (yf-db/open-db)) (yf-db/load-historical! db "AAPL" :period "1y") (yf-db/load-multi-ticker! db ["AAPL" "GOOGL" "MSFT"] :period "1y")

;; Query with SQL (yf-db/query db "SELECT * FROM AAPL ORDER BY timestamp DESC LIMIT 5") (yf-db/query db "SELECT ticker, AVG(close) AS avg_close FROM prices GROUP BY ticker")

;; Close when done (yf-db/close! db)

close! clj

(close! {:keys [db conn]})

Disconnect and close a database opened with open-db.

Example: (close! db)

load-dataset! clj

(load-dataset! {:keys [conn]} dataset & {:keys [table-name] :as opts})

Load an arbitrary tech.v3.dataset into DuckDB under the given table name.

The dataset must have a :name in its metadata, or you must supply :table-name via options. Creates the table and inserts all rows.

Returns the number of rows inserted.

Example: (load-dataset! db my-ds :table-name "prices")

load-historical! clj

(load-historical! {:keys [conn]} ticker & opts)

Fetch historical OHLCV data for ticker and load into DuckDB.

The table name defaults to the ticker symbol (e.g. "AAPL"). Options are passed through to fetch-historical: :period, :interval, :start, :end, :adjusted, :prepost

Returns the number of rows inserted, or nil if fetch fails.

Example: (load-historical! db "AAPL" :period "1y")

load-multi-ticker! clj

(load-multi-ticker! {:keys [conn]}
                    tickers
                    &
                    {:keys [table-name] :or {table-name "prices"} :as opts})

Fetch historical data for multiple tickers and load into DuckDB as a single table.

The combined dataset (with a :ticker column) is loaded under the given table name, defaulting to "prices".

Options are passed through to fetch-historical: :period, :interval, :start, :end, :adjusted, :prepost

Returns the number of rows inserted, or nil if all fetches fail.

Example: (load-multi-ticker! db ["AAPL" "GOOGL" "MSFT"] :period "1y") (query db "SELECT ticker, AVG(close) FROM prices GROUP BY ticker")

open-db clj

(open-db)

(open-db path)

Open an embedded DuckDB database. Pass a file path to persist to disk, or omit for an in-memory database.

Calls duckdb/initialize! automatically on first use.

Returns a map {:db <db> :conn <conn>} that is passed to the other functions.

Examples: (open-db) ; in-memory (open-db "finance.db") ; persistent

query clj

(query {:keys [conn]} sql)

Run a SQL query against the database and return a dataset.

Example: (query db "SELECT * FROM AAPL ORDER BY timestamp DESC LIMIT 10") (query db "SELECT ticker, AVG(close) AS avg_close FROM prices GROUP BY ticker ORDER BY avg_close DESC")

run! clj

(run! {:keys [conn]} sql)

Run a SQL statement ignoring the result. Useful for DDL or DML statements.

Example: (run! db "DROP TABLE IF EXISTS prices")