Open and Plot Parquet Files Online

How to open a Parquet file online

1. 1Open your .parquet file. Drop the file onto the viewer at the top of this page, or click “Choose a file”. Its schema, row count and column statistics appear at once, while the rows are read.
2. 2Look at the schema. Click “Schema” in the file bar for every column’s type, missing values, smallest and largest value and size. The table below the chart scrolls through every row, and you can sort, filter and search it.
3. 3Plot, zoom and compare. The first number column is plotted against the time column straight away. Tick more columns, put one on the right axis, zoom with the scroll wheel, a pinch or a drag, and read each column’s statistics in the side panel.

To try it in Python or another tool too, download the sample Parquet file.

Check this example

“Try a sample Parquet file” opens ten years of made-up hourly UK weather, with London_Temperature, Manchester_Temperature and Edinburgh_Temperature plotted against time. Click Schema in the file bar and it reads 87,672 rows · 20 columns · 10 row groups · SNAPPY, with these rows among the twenty:

Tick Cardiff_Rainfall in the channel list to plot it as well: with the whole file in view, the Statistics panel gives it the same 1,891 missing hours, minimum and maximum, worked out from every row. Read on its own with pyarrow, the Python library for Parquet files, the file gives the same row count, row groups, missing values, minimums and maximums (checked with pyarrow 25.0.1 on 23 October 2026). A test opens the sample on every change to the site and checks that the schema and the Statistics panel show these numbers.

Plot Parquet data, not just read it

A Parquet file viewer that only shows a table tells you the values; a chart shows you what they do. This free Parquet viewer online is also a plotter: every number and date column can be charted, as a line, step, bar, area or scatter chart or a histogram, with a second y axis for a column in other units. Zoom into a moment with the scroll wheel, a pinch or a drag, and the full detail of that range reloads at once, so a one-sample spike in millions of rows stays visible. Add more charts side by side, summarise by month or by category, and save any chart as a PNG image.

A large Parquet viewer, built for big files

Parquet files from data pipelines are often too big for a spreadsheet and slow to open in a notebook. In our tests in Chrome on a recent laptop, a Parquet file with 10 million rows opened and drew its first chart in about 1 second^*.

* Measured in Chrome on a recent laptop (Apple M4): a 141 MB Parquet file (Snappy) with 10,000,000 rows and 5 number columns, from choosing the file to the first chart. Slower devices take longer.

The machine, the browser and how we time a file are on how we measure speed.

Wide files open fast too: number, date and category columns are read with the file, and large text columns, such as free-text notes or IDs, are read the first time you plot them or ask for their statistics. The table shows them straight away. The notes say which columns those are, and “Schema” marks them.

A Parquet schema viewer

The schema is the first thing you see: as soon as you choose a file, the viewer lists every column with its Parquet type, the type it is shown as, its missing values, its smallest and largest value and its size, together with the row count, the row groups and the compression, all taken from the file itself. Afterwards, “Schema” in the file bar shows it again.