From f9fca73d9d4e1403cc76c72c07e3263c460a5f85 Mon Sep 17 00:00:00 2001
From: medusa <newton214@gmail.com>
Date: Tue, 20 Feb 2024 20:41:45 +0000
Subject: [PATCH] Add docs/financial_docs/Database_Schema.md

---
 docs/financial_docs/Database_Schema.md | 75 ++++++++++++++++++++++++++
 1 file changed, 75 insertions(+)
 create mode 100644 docs/financial_docs/Database_Schema.md

diff --git a/docs/financial_docs/Database_Schema.md b/docs/financial_docs/Database_Schema.md
new file mode 100644
index 0000000..b732e27
--- /dev/null
+++ b/docs/financial_docs/Database_Schema.md
@@ -0,0 +1,75 @@
+### Objective
+
+Create a unified database schema to store and analyze forex market data from Oanda, focusing on multiple currency pairs with the flexibility to support a wide range of analytical and machine learning workloads.
+
+### Schema Design
+
+The schema is designed to store time-series data for various forex instruments, capturing price movements and trading volumes over time, along with allowing for the storage of additional, flexible data points.
+
+#### Proposed Schema for SQLite3
+
+```sql
+CREATE TABLE forex_data (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    instrument TEXT NOT NULL,
+    timestamp DATETIME NOT NULL,
+    open REAL NOT NULL,
+    high REAL NOT NULL,
+    low REAL NOT NULL,
+    close REAL NOT NULL,
+    volume INTEGER,
+    additional_info TEXT
+);
+```
+
+#### Adaptation for TimescaleDB (PostgreSQL)
+
+```sql
+CREATE TABLE forex_data (
+    id SERIAL PRIMARY KEY,
+    instrument VARCHAR(10) NOT NULL,
+    timestamp TIMESTAMPTZ NOT NULL,
+    open NUMERIC NOT NULL,
+    high NUMERIC NOT NULL,
+    low NUMERIC NOT NULL,
+    close NUMERIC NOT NULL,
+    volume NUMERIC,
+    additional_info JSONB,
+    CONSTRAINT unique_instrument_timestamp UNIQUE (instrument, timestamp)
+);
+```
+
+### Key Components Explained
+
+- **id**: A unique identifier for each row. Simplifies data retrieval and management, especially for ML applications where each data point might need to be uniquely identified.
+
+- **instrument**: Specifies the forex pair (e.g., 'EUR_USD', 'GBP_JPY'), allowing data from multiple instruments to be stored in the same table.
+
+- **timestamp**: Records the datetime for each data point. It's crucial for time series analysis. `TIMESTAMPTZ` in TimescaleDB ensures time zone awareness.
+
+- **open, high, low, close**: Represent the opening, highest, lowest, and closing prices for the instrument within the specified time interval.
+
+- **volume**: Represents the trading volume. It's optional, recognizing that volume data might not always be available or relevant.
+
+- **additional_info**: A flexible JSONB (or TEXT in SQLite) column for storing any additional structured data related to the data point, such as bid/ask prices, computed indicators, or metadata.
+
+- **unique_instrument_timestamp**: Ensures data integrity by preventing duplicate entries for the same instrument and timestamp.
+
+### Transitioning from SQLite3 to TimescaleDB
+
+This schema is designed with compatibility in mind. The transition from SQLite3 to TimescaleDB involves type adjustments and taking advantage of TimescaleDB's features for time-series data. Upon migration, you would:
+
+1. Convert data types where necessary (e.g., `TEXT` to `VARCHAR`, `DATETIME` to `TIMESTAMPTZ`, `TEXT` containing JSON to `JSONB`).
+2. Apply TimescaleDB's time-series optimizations, such as creating a hypertable for efficient data storage and querying.
+
+### Documentation and Usage Notes
+
+- **Granularity**: Decide on the granularity (e.g., tick, minute, hourly, daily) based on your analytical needs. This affects the `timestamp` and potentially the `volume` and price precision.
+- **Time Zone Handling**: Be mindful of time zones, especially if analyzing global markets. `TIMESTAMPTZ` in TimescaleDB helps manage time zone complexities.
+- **Data Integrity**: The unique constraint on `instrument` and `timestamp` prevents data duplication, ensuring the database's reliability for analysis.
+- **Extensibility**: The `additional_info` JSONB column allows for the addition of new data points without schema modifications, offering extensibility for future analysis needs.
+- **Machine Learning and Analysis**: This schema supports direct use with Python's data analysis libraries (e.g., Pandas for data manipulation, Scikit-learn for ML modeling) by facilitating the extraction of features directly from stored data.
+
+### Conclusion
+
+This guide provides a blueprint for a database schema capable of supporting comprehensive forex data analysis and machine learning workloads, from initial development with SQLite3 to a scalable, production-ready setup with TimescaleDB. By focusing on flexibility, scalability, and compatibility, this schema ensures that your database can grow and evolve alongside your analytical capabilities, providing a solid foundation for extracting insights from forex market data.
\ No newline at end of file