Opening a Parquet, SQLite or Excel data export

A data team, an app or a reporting system hands you a .parquet, .db or .xlsx file, and you only want to know what’s in it, or answer one question from it. You don’t need Python, a database or a spreadsheet licence: these tools open the files in your browser, and nothing is uploaded.

1. Open it in the right tool

Go by the file’s name, or by File Type if the name doesn’t say:

• .parquet, written by pandas, Polars, Spark, DuckDB and most data lakes: the Parquet Viewer.
• .db, .sqlite or .sqlite3, the database inside many apps: the SQLite Viewer.
• .xlsx, an Excel workbook: CSV ⇄ JSON, which reads each sheet as CSV.

2. Check its columns and types

Before trusting the numbers, check what each column holds:

• In the Parquet Viewer, Schema lists every column with its type, such as TIMESTAMP (ms, UTC) or DECIMAL(10, 2), with lists and structs indented under their parent. Column sizes gives each column’s nulls, which show at a glance which columns are mostly empty.
• In the SQLite Viewer, Table lists the tables with their row counts, and Schema has the CREATE statements with each column’s type.
• An Excel sheet has no types: in CSV ⇄ JSON every value is text, so check that dates and numbers came through as you expect.

3. Look through the rows

The Parquet Viewer shows 1,000 rows at a time: type a row number to jump further in. Choose a column’s name to sort the rows shown, and type in Filter these rows to keep those containing some text. In CSV ⇄ JSON, choose Show as a table for the same sorting and filtering. In the SQLite Viewer, choosing a table shows its first 1,000 rows.

4. Ask it a question with SQL

Sorting and filtering only go so far. To count, total or group the rows, or match them against another file, use SQL in the SQLite Viewer:

• An SQLite database is ready to query: type the SQL and press Run.
• For a Parquet file, choose Save every row as CSV in the Parquet Viewer, then open that CSV in the SQLite Viewer. It becomes a table named after the file.
• For an Excel sheet, open it in CSV ⇄ JSON and choose To SQL: copy the statements it writes into the SQLite Viewer’s SQL box and press Run, and the sheet becomes a table named data.

Several CSV files become several tables, so you can JOIN them. For example, with a file orders.csv:

SELECT country, COUNT(*) AS orders, SUM(total) AS value
FROM orders GROUP BY country ORDER BY value DESC;

5. Save or chart the answer

Save result as CSV in the SQLite Viewer keeps the answer for a spreadsheet or another system. For a quick picture, Copy result as CSV and paste it into the Chart Maker: the first column gives the labels, the header row names the series. To hand someone a spreadsheet instead, paste the CSV into CSV ⇄ JSON and choose Save as Excel (.xlsx).

If the export itself is a messy CSV, with broken rows or garbled names, see From a messy CSV to clean data.