Wrds Data Access

Background & Key Concepts

WRDS Architecture

WRDS (Wharton Research Data Services) exposes PostgreSQL databases via:

1. Python package (wrds): Connects via SSH tunnel + PostgreSQL; requires institutional subscription
2. SQL queries: Standard SQL on schemas like comp (Compustat), crsp, taq

Key Identifiers

CRSP-Compustat Merge (CCM)

The crsp.ccmxpf_lnkhist table maps gvkey → permno with date-valid links. Always filter by linktype IN ('LC','LU','LX','LD') and linkprim IN ('P','C').

Environment Setup

Install Dependencies

pip install wrds>=3.1 pandas>=2.0 numpy>=1.24 sqlalchemy>=2.0 matplotlib>=3.7

WRDS Credentials

# Credentials are stored securely in ~/.pgpass after first login
# Set environment variables (do NOT hardcode)
export WRDS_USERNAME="<your-wrds-username>"

Connect to WRDS

import os
import wrds

# Credentials via environment variable
wrds_username = os.getenv("WRDS_USERNAME", "")
if not wrds_username:
    raise ValueError("Set WRDS_USERNAME environment variable")

db = wrds.Connection(wrds_username=wrds_username)
print("Connected to WRDS")

# List available libraries
libs = db.list_libraries()
print("Available libraries (first 10):", libs[:10])

Core Workflow

Step 1: Compustat Annual Fundamentals

import os
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

# ------------------------------------------------------------------ #
# Download Compustat annual fundamentals for S&P 500 firms
# Key variables for accounting research
# ------------------------------------------------------------------ #