Run3 sec 13f dataset structure

SEC 13F Dataset Structure

The datasets are stored in /root/2025-q2 and /root/2025-q3 folders.

Check for Scripts Directory

The dataset likely comes with pre-built Python scripts. Check:

• /root/2025-q3/scripts/ or /root/2025-q2/scripts/
• Common scripts: search_fund.py, one_fund_analysis.py, holding_analysis.py, search_stock_cusip.py

# First, explore the directory structure
find /root/2025-q2 -type f | head -40
find /root/2025-q3 -type f | head -40
ls -la /root/2025-q2/scripts/ 2>/dev/null
ls -la /root/2025-q3/scripts/ 2>/dev/null
ls -la /root/2025-q2/*.py 2>/dev/null
ls -la /root/2025-q3/*.py 2>/dev/null

Always check the scripts first and read their source code to understand how they handle data parsing, filtering, and deduplication. These scripts handle edge cases correctly.

Data Files (TSV format)

Key TSV files typically found:

• COVERPAGE.tsv — Fund cover page info (filing metadata, accession numbers)
• INFOTABLE.tsv — Holdings data (stocks, CUSIPs, values, shares)
• SUMMARYPAGE.tsv — Summary info
• Other supporting files

Key Columns in INFOTABLE.tsv

• ACCESSION_NUMBER — Links to COVERPAGE
• NAMEOFISSUER — Company name
• CUSIP — Stock identifier (9-character)
• VALUE — Value in thousands of dollars
• SSHPRNAMT — Number of shares/principal amount
• SSHPRNAMTTYPE — Type: "SH" for shares, "PRN" for principal
• PUTCALL — If populated with "PUT" or "CALL", it's an option, not a stock
• INVESTMENTDISCRETION — Type of discretion

Critical: PUTCALL Column Values

The PUTCALL column may contain unexpected values. Always inspect:

import pandas as pd
info = pd.read_csv('/root/2025-q3/INFOTABLE.tsv', sep='\t')
print(info['PUTCALL'].unique())
print(info['PUTCALL'].value_counts(dropna=False))