Skip to content
xbbg logo xbbg

Python

xbbg is the primary public package in this monorepo and the modern standard Bloomberg API for Python. It keeps the familiar Bloomberg-oriented Python API while moving transport, session management, authentication, and Arrow-native data handling into a native Rust engine backed by Bloomberg’s C/C++ SDK. It is not a thin homegrown wrapper around the blpapi Python package; it is designed to replace direct blpapi usage for almost every normal Bloomberg application workflow.

What you get

Section titled “What you get”

Enterprise package

Install xbbg from PyPI for notebooks, research workflows, and enterprise Python services that need ZFP, B-PIPE auth, TLS, failover, SOCKS5, SDK logging, and diagnostics.

Backend choice

Return data as a Narwhals DataFrame by default, preferring a real PyArrow table when PyArrow is installed and falling back to xbbg’s native Arrow carrier for minimal installs. Explicitly convert to native, PyArrow, pandas, Polars, DuckDB, and other supported backends when your pipeline needs them.

Async support

Use the synchronous API in scripts or the a-prefixed functions inside async applications without switching packages.

Shared engine

Python gets the same Rust-backed request engine, validation model, Arrow transport, session pooling, and Bloomberg C++ SDK integration used across the public runtime surface.

Why xbbg is different

Section titled “Why xbbg is different”

Most Python Bloomberg packages are simple wrappers over Bloomberg’s blpapi Python bindings. That path is fine for small scripts, but it leaves session lifecycle, event parsing, concurrency, diagnostics, enterprise middleware, and enterprise connectivity in Python glue code. xbbg sits lower and more professionally in the stack: the Python API calls a native Rust engine that binds to Bloomberg’s C/C++ SDK, manages sessions and worker pools, builds typed Arrow output, and exposes enterprise transports such as ZFP over leased lines, B-PIPE/SAPI auth, TLS, failover servers, SOCKS5, SDK logging, validation, request diagnostics, and request middleware.

That architecture is why xbbg can be both simple at the call site (xbbg.bdp(...)) and serious underneath: near-complete blpapi workflow coverage, generic service requests for uncommon operations, async execution, reusable sessions, zero-copy-oriented Arrow handoff, enterprise middleware for audit/logging/metrics/policy hooks, benchmark coverage across changes, and multiple dataframe backends without switching libraries.

Compared with other Bloomberg Python wrappers

Section titled “Compared with other Bloomberg Python wrappers”

xbbg is not just another pandas wrapper around Bloomberg’s blpapi Python package. It is a native Bloomberg stack: Python ergonomics on top of a Rust engine backed by Bloomberg’s C/C++ SDK. That distinction matters once a workflow grows beyond a single bdp() call.

  • Raw blpapi is official but low-level: every application has to rebuild session startup, service opening, request construction, correlation IDs, event-loop parsing, error mapping, retries, logging, and dataframe conversion. xbbg keeps the SDK power underneath while replacing that boilerplate with typed helpers, generic service requests, structured errors, async execution, and Arrow-native output.
  • bbg-fetch / BloombergFetch is a small homegrown pandas helper package. It targets Python 3.9–3.12, installs NumPy and pandas as hard dependencies, and does not cover intraday bars, ticks, streaming, BQL, BEQS, BSRCH, BQR, BTA, ZFP, B-PIPE/SAPI auth, SDK logging, multi-backend output, typed Arrow transport, async execution, or generic Bloomberg service requests.
  • pdblp is legacy and explicitly superseded by blp; it has no modern Requires-Python floor, depends on old pandas-era assumptions, and is not under active development.
  • blp is cleaner than pdblp but still sits in the classic Python/pandas wrapper category (python>=3.6, mandatory pandas). xbbg goes lower in the stack with native Rust execution, reusable worker pools, Arrow output, enterprise transports, and near-complete blpapi workflow coverage.
  • polars-bloomberg locks the workflow into Polars and has useful but partial BQL/search-style coverage. It is not full Bloomberg API coverage and does not replace xbbg’s BDP/BDS/BDH/BDIB/BDTICK/streaming surface, enterprise transports, middleware, diagnostics, generic service requests, or multi-backend architecture.

xbbg is also the only package in this comparison with a real CI matrix across every supported Python version: 3.10, 3.11, 3.12, 3.13, and 3.14.

Enterprise middleware

Section titled “Enterprise middleware”

xbbg includes request middleware for teams that need platform-wide controls around Bloomberg access. Middleware receives request context and environment snapshots, so one shared pipeline can add audit logging, request IDs, metrics, tracing, entitlement checks, policy enforcement, standardized error handling, and compliance hooks without wrapping every individual bdp(), bdh(), bql(), or streaming call.

Quick example

Section titled “Quick example”
import xbbg


xbbg.configure(host='localhost', port=8194)
prices = xbbg.bdp(
    ['AAPL US Equity', 'MSFT US Equity'],
    ['PX_LAST', 'VOLUME'],
)

Prefer the top-level helpers in these docs (xbbg.bdp(...), xbbg.configure(...)). If you are coming from the repository README, the module-style import remains supported too: from xbbg import blp and blp.bdp(...).

Next steps

Section titled “Next steps”