Historical Data

ThetaDataDx provides a generated historical data surface across five asset categories. All historical data is accessed through the unified ThetaDataDxClient, which communicates over gRPC with ThetaData's MDDS servers. gRPC, protobuf parsing, zstd decompression, and FIT decoding run inside the thetadatadx Rust crate, regardless of which language binding you call.

Connecting

Rust

use thetadatadx::{ThetaDataDxClient, Credentials, DirectConfig};

let creds = Credentials::from_file("creds.txt")?;
let client = ThetaDataDxClient::connect(&creds, DirectConfig::production()).await?;

Python

from thetadatadx import Credentials, Config, ThetaDataDxClient

creds = Credentials.from_file("creds.txt")
client = ThetaDataDxClient(creds, Config.production())

TypeScript

import { ThetaDataDxClient } from 'thetadatadx';

const client = await ThetaDataDxClient.connectFromFile('creds.txt');

C++

auto creds = tdx::Credentials::from_file("creds.txt");
auto client = tdx::UnifiedClient::connect(creds, tdx::Config::production());

Endpoint Categories

Category	Endpoints	Page
Stocks	14 endpoints - list, snapshots, history, at-time, streaming	Stock Endpoints
Options	34 endpoints - list, snapshots, history, Greeks, trade Greeks, at-time	Option Endpoints
Indices	9 endpoints - list, snapshots, history, at-time	Index Endpoints
Calendar	3 endpoints - trading schedule, holidays, early closes	Calendar
Rates	1 endpoint - interest rate EOD history	Rates

Date Format

All dates are YYYYMMDD strings: "20240315" for March 15, 2024.

Interval Format

Intervals are millisecond strings: "60000" for 1 minute, "300000" for 5 minutes, "3600000" for 1 hour.

DataFrame Support (Python)

Every historical method returns a typed list wrapper (EodTickList, OhlcTickList, StringList, ...). Chain .to_polars() / .to_pandas() / .to_arrow() / .to_list() on the return value for the matching representation:

df  = tdx.stock_history_eod("AAPL", "20240101", "20240301").to_polars()
pdf = tdx.stock_history_eod("AAPL", "20240101", "20240301").to_pandas()
tbl = tdx.stock_history_eod("AAPL", "20240101", "20240301").to_arrow()
lst = tdx.stock_history_eod("AAPL", "20240101", "20240301").to_list()

The list wrapper itself behaves like a Python sequence (len(ticks), ticks[i], for t in ticks:), so most callers skip the terminal entirely.

Requires pip install thetadatadx[pandas] or [polars] / [arrow] depending on the terminal.

Time Reference

Time (ET)	time_of_day
9:30 AM	09:30:00.000
12:00 PM	12:00:00.000
4:00 PM	16:00:00.000

Empty Responses

When a query returns no data (e.g., a non-trading date), the SDK returns an empty collection rather than an error. Check for emptiness using the appropriate idiom for your language:

Rust

if eod.is_empty() {
    println!("No data for this date range");
}

Python

if not eod:
    print("No data for this date range")

TypeScript

if (eod.length === 0) {
    console.log('No data for this date range');
}

C++

if (eod.empty()) {
    std::cout << "No data for this date range" << std::endl;
}