When dealing with repetitive document creation, using a template-based approach significantly reduces manual effort. The Python package docxtpl lets you load a .docx template, replace placeholders with real values, and save the result. Combined with pandas for reading tabular data, you can bulk-generate personalized documents from a single template.

Installation

pip install docxtpl pandas

Example: Generating Admission Letters

Assume you have a Word template (admission_template.docx) with placeholders wrapped in double curly braces, such as {{name}}, {{faculty}}, {{major}}, {{start_date}}, and {{end_date}}.

Your data source is a CSV file (candidates.csv) structured like this:

name	faculty	major	start_date	end_date
Alice	Computer Science	AI	2025-09-01	2029-06-30
Bob	Computer Science	AI	2025-09-01	2029-06-30

Step 1 – Import Libraries

from docxtpl import DocxTemplate
import pandas as pd
import os

Step 2 – Prepare Output Directory

base_dir = os.getcwd()
out_dir = os.path.join(base_dir, 'letters_output')
os.makedirs(out_dir, exist_ok=True)

Step 3 – Read Data from CSV

df = pd.read_csv(
    os.path.join(base_dir, 'candidates.csv'),
    encoding='utf-8'
)

# Strip whitespace from all object columns
df = df.apply(lambda col: col.str.strip() if col.dtype == 'object' else col)

Step 4 – Generate Documents in a Loop

template_path = os.path.join(base_dir, 'admission_template.docx')

# Load the template once outside the loop
tpl = DocxTemplate(template_path)

for idx, row in df.iterrows():
    # Build context dictionary from the current row
    context = {
        'name': row['name'],
        'faculty': row['faculty'],
        'major': row['major'],
        'start_date': row['start_date'],
        'end_date': row['end_date']
    }

    # Render the template with current context
    tpl.render(context)

    # Construct output filename
    save_name = f"{row['name']}_admission_letter.docx"
    out_path = os.path.join(out_dir, save_name)

    # Save the generated document
    tpl.save(out_path)

The script iterates over each row, fills the template, and saves a personalized Word file into letters_output/. The template is loaded once; subsequent render() and save() calls reuse the same instance, which is efficient for large batches.

Important Considerations

• Encoding issues: If you encounter garbled characters, adjust the encoding parameter (e.g., 'gbk' or 'utf-8').
• Whitespace in data: Use .str.strip() on string columns to prevent unexpected line breaks in the output.
• File permissions: Insure the CSV file is not open in another program during execution.
• Template re-selection: If you see the same content in all output files, verify that you are not reloading the template inside the loop (loading after each save resets it).