Extract tables from PDFs with Python

Extracting tables from PDFs using Python can be efficiently done with several libraries designed for PDF parsing and data extraction. Here’s a detailed guide covering the most popular tools and methods to extract tables from PDFs, along with code examples.

1. Using Tabula-py

Tabula-py is a Python wrapper for Tabula, a Java library that extracts tables from PDFs into DataFrames.

Installation:

bash
CopyEdit
pip install tabula-py

You also need to have Java installed on your system because Tabula depends on it.

Basic usage:

python
CopyEdit
import tabula

# Path to the PDF file
file_path = "sample.pdf"

# Extract all tables from the PDF as a list of DataFrames
tables = tabula.read_pdf(file_path, pages='all', multiple_tables=True)

for i, table in enumerate(tables):
    print(f"Table {i+1}")
    print(table)

Save extracted tables as CSV:

python
CopyEdit
tabula.convert_into(file_path, "output.csv", output_format="csv", pages='all')

2. Using Camelot

Camelot works best with PDFs that have well-defined tables with clear borders.

Installation:

bash
CopyEdit
pip install camelot-py[cv]

Note: Camelot requires Ghostscript and Tkinter installed on your system.

Basic usage:

python
CopyEdit
import camelot

# Read tables from PDF
tables = camelot.read_pdf('sample.pdf', pages='1')

# Number of tables found
print(f"Total tables extracted: {tables.n}")

# Export tables to CSV or JSON
tables[0].to_csv('table1.csv')
tables[0].to_json('table1.json')

# Print first table as DataFrame
print(tables[0].df)

Camelot also supports two parsing methods:

• stream (for tables without borders)

• lattice (for tables with borders)

Example:

python
CopyEdit
tables = camelot.read_pdf('sample.pdf', flavor='stream', pages='1')

3. Using PyMuPDF (fitz) and Pandas for Custom Extraction

PyMuPDF can extract raw text, but does not directly extract tables. However, if tables are simple, you can extract text and use string processing to parse tables.

Installation:

bash
CopyEdit
pip install pymupdf

Example to extract text:

python
CopyEdit
import fitz  # PyMuPDF

file_path = "sample.pdf"
doc = fitz.open(file_path)

for page_num in range(len(doc)):
    page = doc.load_page(page_num)
    text = page.get_text()
    print(text)

From the extracted text, use Python string methods or regex to convert to table format.

4. Using pdfplumber

pdfplumber is a powerful tool for extracting tables from PDFs, even for complex layouts.

Installation:

bash
CopyEdit
pip install pdfplumber

Extract tables example:

python
CopyEdit
import pdfplumber

with pdfplumber.open("sample.pdf") as pdf:
    first_page = pdf.pages[0]
    tables = first_page.extract_tables()
    
    for table in tables:
        for row in table:
            print(row)

You can convert extracted tables into pandas DataFrames:

python
CopyEdit
import pandas as pd

with pdfplumber.open("sample.pdf") as pdf:
    first_page = pdf.pages[0]
    tables = first_page.extract_tables()
    
    for i, table in enumerate(tables):
        df = pd.DataFrame(table[1:], columns=table[0])
        print(f"Table {i+1}")
        print(df)

5. Summary of Libraries and Use Cases

Tips for Better Table Extraction

• Check if the PDF contains actual tables or just images/scans. For scanned PDFs, consider OCR libraries like Tesseract.

• Choose the right parsing flavor (Camelot’s lattice vs stream).

• Preprocess PDFs if needed (e.g., crop or convert pages).

• Test extraction on a sample page before running on the entire document.

• Handle merged cells and multi-line headers carefully when converting tables.

With these methods and tools, extracting tables from PDFs using Python becomes manageable and customizable for various types of documents.