Parquet Visualizer

Inspect and query very large parquet files fast

sql

What's Parquet Visualizer

Parquet Visualizer is a tool that helps you easily query with SQL and inspect very large parquet files fast and easy.

What can you do with Parquet Visualizer?

Inspect Data

You can inspect the binary data of parquet files in a human readable tabular format with pagination. You can also change the page size.

data

Run SQL Queries on a Parquet File

You can query a parquet file with DuckDB SQL. You can also search within the result, paginate the result or change the page size.

sql

Search within rows of Query Result

By typing free text in the search box, find specific values of the query result.

Export Query Result as CSV, JSON, ndJSON or Parquet to Disk

By clicking on the export button in the query tab, you can save your query result in CSV, JSON, ndJSON or Parquet format to disk.

Copy Query Result to Clipboard

By clicking on the copy button in the querytab, you can copy the query result data to the clipboard.

Inspect Struct Value

You can easily inspect complex struct values by clicking on the cell, which shows a popup containing the value of the struct.

complex

Inspect Schema

You can inspect the schema of the parquet file by clicking on the Schema tab, in which you can paginate if the file has many columns.

You can also inspect the struct type by clicking on the cell, which will show a popup containing the struct data type.

schema

Inspect Metadata

You can inspect the metadata in tabular format by clicking on the Metadata tab.

metadata

Configuration

The following configuration options are available:

name	default	description
`parquet-visualizer.backend`	`duckdb`	Backend for reading the parquet file. Options: `duckdb`, `parquet-wasm`
`parquet-visualizer.defaultPageSizes`	`[20, 50, 100, 500]`	Set the default page size for data and query tab.
`parquet-visualizer.defaultQuery`	`SELECT *\r\nFROM data\r\nLIMIT 1000;`	Default SQL query for parquet file. The table `data` should remain the same.

Parquet backends

This extension supports two different types of backends for visualizing and querying parquet files.

DuckDB

DuckDB is the primary backend used for uncompressed and compressed parquet files (except for the BROTLI compression codec.)

Parquet-wasm

parquet-wasm is a backend that uses a Rust implementation of arrow and parquet. It supports all compression codecs except LZ4.

Parquet Visualizer

Lucien Martijn

Parquet Visualizer

Inspect and query very large parquet files fast

What's Parquet Visualizer

What can you do with Parquet Visualizer?

Inspect Data

Run SQL Queries on a Parquet File

Search within rows of Query Result

Export Query Result as CSV, JSON, ndJSON or Parquet to Disk

Copy Query Result to Clipboard

Inspect Struct Value

Inspect Schema

Inspect Metadata

Configuration

Parquet backends

DuckDB

Parquet-wasm

Frontend