polars-bloomberg 0.2.3

Python library providing a Polars DataFrame interface for easy and intuitive access to Bloomberg API.

Polars + Bloomberg Open API

Polars + Bloomberg Open API is a Python library that facilitates integration of Bloomberg data into Polars DataFrames. Designed for users familiar with Pandas or Excel, it offers minimal-boilerplate functions such as bdp(), bdh(), and bql(). Leverage Polars' high-performance capabilities alongside the Bloomberg API for lightning-fast DataFrame operations and a minimal memory footprint.

Key Benefits:

• Intuitive "Excel-like" methods: bdp(), bdh(), bql()
• Outputs data as Polars DataFrames
• Lightweight design with no dependency on pandas
• Quickly prototype, test, and scale complex financial analyses.

Prerequisites

• Bloomberg Access: A valid Bloomberg terminal license.
• Bloomberg Python API: The blpapi library must be installed. See the Bloomberg API Library for guidance.
• Python Version: Python 3.8+ recommended.
• Installation: pip install polars-bloomberg

Quick Start Guide (5 Minutes)

Below is a simple example to get you started. For more comprehensive examples, please see the examples/ directory.

Concept:
BQuery is your main interface. Using a context manager ensures the connection opens and closes cleanly. Within this session, you can use:

• bq.bdp() for Bloomberg Data Points (single-value fields).
• bq.bdh() for Historical Data (time series).
• bq.bql() for complex Bloomberg Query Language requests.

BDP - Bloomberg Data Point

Example: Fetching the Last Price of Apple and Microsoft

from polars_bloomberg import BQuery

with BQuery() as bq:
    df = bq.bdp(['AAPL US Equity', 'MSFT US Equity'], ['PX_LAST'])

shape: (2, 2)

security	PX_LAST
str	f64
"AAPL US Equity"	242.84
"MSFT US Equity"	443.57

More BDP Examples

BDP with different column types

polars-bloomberg correctly infers column type as shown in this example:

with BQuery() as bq:
    df = bq.bdp(["XS2930103580 Corp", "USX60003AC87 Corp"],
                ["SECURITY_DES", "YAS_ZSPREAD", "CRNCY", "NXT_CALL_DT"])

shape: (2, 5)

security	SECURITY_DES	YAS_ZSPREAD	CRNCY	NXT_CALL_DT
str	str	f64	str	date
"XS2930103580 Corp"	"SEB 6 3/4 PERP"	327.309349	"USD"	2031-11-04
"USX60003AC87 Corp"	"NDAFH 6.3 PERP"	315.539222	"USD"	2031-09-25

BDP with overrides

User can submit list of tuples with overrides

with BQuery() as bq:
    df = bq.bdp(["IBM US Equity"], ["PX_LAST", "CRNCY_ADJ_PX_LAST"], 
                overrides=[("EQY_FUND_CRNCY", "SEK")])

</style> shape: (1, 3)

security	PX_LAST	CRNCY_ADJ_PX_LAST
str	f64	f64
"IBM US Equity"	238.04	2607.401

BDP with date overrides

Overrides for dates has to be in format YYYYMMDD

with BQuery() as bq:
    df = bq.bdp(["USX60003AC87 Corp"], ["SETTLE_DT"], overrides=[("USER_LOCAL_TRADE_DATE", "20241014")])

shape: (1, 2)

security	SETTLE_DT
str	date
"USX60003AC87 Corp"	2024-10-15

with BQuery() as bq:
    df = bq.bdp(['USDSEK Curncy', 'SEKCZK Curncy'], 
                ['SETTLE_DT', 'PX_LAST'], 
                overrides=[('REFERENCE_DATE', '20200715')]
               )

shape: (2, 3)

security	SETTLE_DT	PX_LAST
str	date	f64
"USDSEK Curncy"	2020-07-17	10.9343
"SEKCZK Curncy"	2020-07-17	2.1718

BDH - Bloomberg Data History

with BQuery() as bq:
    df = bq.bdh(['AAPL US Equity', 'TSLA US Equity'], 
                ['PX_LAST', 'VOLUME'], 
                start_date=date(2019, 1, 1), 
                end_date=date(2019, 1, 10))

shape: (14, 4)

security	date	PX_LAST	VOLUME
str	date	f64	f64
"AAPL US Equity"	2019-01-02	39.48	1.48158948e8
"AAPL US Equity"	2019-01-03	35.548	3.6524878e8
…	…	…	…
"TSLA US Equity"	2019-01-09	22.5687	8.1494175e7
"TSLA US Equity"	2019-01-10	22.998	9.084531e7

More BDH Examples

BDH with options - periodicitySelection: Monthly

with BQuery() as bq:
    df = bq.bdh(['AAPL US Equity'], 
                ['PX_LAST'], 
                start_date=date(2019, 1, 1), 
                end_date=date(2019, 3, 29),
                options={"periodicitySelection": "MONTHLY"})

shape: (3, 3)

security	date	PX_LAST
str	date	f64
"AAPL US Equity"	2019-01-31	41.61
"AAPL US Equity"	2019-02-28	43.288
"AAPL US Equity"	2019-03-29	47.488

BQL - Bloomberg Query Language

Allows to run complex bql queries and get result in wide polars.DataFramewith correct polars types

with BQuery() as bq:
    df = bq.bql("get(px_last) for(['IBM US Equity', 'OMX Index'])")

shape: (2, 4)

ID	px_last	px_last.DATE	px_last.CURRENCY
str	f64	date	str
"IBM US Equity"	238.04	2024-12-07	"USD"
"OMX Index"	2614.268	2024-12-07	"SEK"

More BQL Examples

Actual and Forward EPS Estimates

df = bq.bql("""
    let(#eps=is_eps(fa_period_type='A',
                    fa_period_offset=range(-4,2));)
    get(#eps)
    for(['IBM US Equity'])
""")

shape: (7, 6)

ID	#eps	#eps.REVISION_DATE	#eps.AS_OF_DATE	#eps.PERIOD_END_DATE	#eps.CURRENCY
str	f64	date	date	date	str
"IBM US Equity"	10.63	2022-02-22	2024-12-07	2019-12-31	"USD"
"IBM US Equity"	6.28	2023-02-28	2024-12-07	2020-12-31	"USD"
…	…	…	…	…	…
"IBM US Equity"	9.236	2024-12-07	2024-12-07	2025-12-31	"USD"

Average issuer OAS spread per maturity bucket

query = """
let( 
    #bins = bins(maturity_years,
                 [3,9,18,30],
                 ['(1) 0-3','(2) 3-9','(3) 9-18','(4) 18-30','(5) 30+']);
    #average_spread = avg(group(spread(st=oas),#bins));
)
get(#average_spread)
for(filter(bonds('NVDA US Equity', issuedby = 'ENTITY'),
           maturity_years != NA))
"""

with BQuery() as bq:
    df = bq.bql(query)

shape: (5, 5)

ID	#average_spread	#average_spread.DATE	#average_spread.ORIG_IDS	#average_spread.#BINS
str	f64	date	str	str
"(1) 0-3"	30.74	2024-12-08	"QZ552396 Corp"	"(1) 0-3"
"(2) 3-9"	59.79	2024-12-08	null	"(2) 3-9"
"(3) 9-18"	105.39	2024-12-08	"BH393780 Corp"	"(3) 9-18"
"(4) 18-30"	131.72	2024-12-08	"BH393781 Corp"	"(4) 18-30"
"(5) 30+"	150.33	2024-12-08	"BH393782 Corp"	"(5) 30+"

Technical Analysis: stocks with 20d EMA > 200d EMA and RSI > 70

with BQuery() as bq:
    df = bq.bql(
        """
        let(#ema20=emavg(period=20); 
            #ema200=emavg(period=200); 
            #rsi=rsi(close=px_last());)
        get(name(), #ema20, #ema200, #rsi)
        for(filter(members('OMX Index'), 
                    and(#ema20 > #ema200, #rsi > 70)))
        with(fill=PREV)
        """
    )

shape: (2, 10)

ID	name()	#ema20	#ema20.DATE	#ema20.CURRENCY	#ema200	#ema200.DATE	#ema200.CURRENCY	#rsi	#rsi.DATE
str	str	f64	date	str	f64	date	str	f64	date
"SKFB SS Equity"	"SKF AB"	210.185019	2024-12-08	"SEK"	204.16756	2024-12-08	"SEK"	72.255568	2024-12-08
"ABB SS Equity"	"ABB Ltd"	623.496942	2024-12-08	"SEK"	561.902577	2024-12-08	"SEK"	72.144556	2024-12-08

API Documentation

Read the API documentation in examples/ directory

More Examples

Explore additional usage examples in the examples/ directory.

Bloomberg Documentation

For documentation on the Bloomberg API, check out the Bloomberg Developer's page.