A Python script that turns your CSV files into JSON or XML in the exact structure your system expects

Example result

python merge_sales.py --month 2026-10

2026-11-02 09:14:03 INFO  Folder: sales/2026-10, 9 files found
2026-11-02 09:14:03 INFO  shop-01.csv: 1,284 rows
2026-11-02 09:14:04 INFO  shop-02.xlsx: 967 rows
2026-11-02 09:14:04 WARN  shop-07.csv: 3 rows without a price skipped (lines 88, 412, 913)
2026-11-02 09:14:05 INFO  Done: 10,412 rows -> report-2026-10.xlsx (sheets "Sales" and "Totals")

[columns]
date = Date, Sale date, Day
product = Product, Item, Name
quantity = Qty, Quantity, Pcs
price = Price, Unit price

def read_shop_file(path: Path, columns: dict[str, list[str]]) -> pd.DataFrame:
    """Reads one shop's file and gives it the report's column names."""
    table = pd.read_excel(path) if path.suffix == ".xlsx" else pd.read_csv(path, encoding="utf-8-sig")
    renamed = {}
    for name, variants in columns.items():
        found = next((c for c in table.columns if c.strip() in variants), None)
        if found is None:
            raise ReportError(f"{path.name}: no column for '{name}' (looked for {', '.join(variants)})")
        renamed[found] = name
    table = table.rename(columns=renamed)[list(columns)]
    table["shop"] = path.stem
    return table

About this service

Many systems take data only as JSON or XML: an import into a shop or a CRM, a partner's upload, a feed for an app. This Python script reads your CSV export and writes the file in the structure you need: the field names, which fields sit inside others, numbers and dates as the right types, and the encoding. In Premium it can also group rows into one record, such as an order with all its items. Send us a few rows of the CSV and an example or description of the JSON or XML the other side wants (its documentation or a sample file). You run the script yourself on your computer. It converts the data you give it: it does not upload the result anywhere or check it against the other system.

What is included

• What you getA ZIP archive with the Python script, a README with step-by-step instructions and, from Standard, a settings file with the column-to-field mapping; the code is also shown on the order page.
• JSON or XMLOr the two side by side from one CSV
• Your structureField names and nesting as in your sample or documentation
• Right typesNumbers, true or false, dates in the format you need
• Encodings handledUTF-8 output from CSV in any common encoding
• Readable codeWith comments, easy to add a field later

How it works

1. 1
   SamplesYou send a few CSV rows and a sample of the JSON or XML you need
2. 2
   CodeWe write the script and the guide to running it
3. 3
   Try itYou convert your file and try it in the system that needs it
4. 4
   ChangesYou ask for changes within your package and send any error message

What we need from you

10-20 rows of your CSV (made-up values are fine), and what the result must look like: a sample JSON or XML file, the part of the other system's documentation that describes it, or an XSD schema if there is one. Which column goes to which field, how dates and numbers should be written, and whether several rows make one record. Your system: Windows, macOS or Linux.

We will ask you:

• Where will the JSON or XML go?
• What does the CSV look like?
• What must the result look like?
• Where will it run?

Compare packages

Questions and answers

Did not find your answer? Ask us