devinrsmith / deephaven-parquet-viewer

A browser-based Parquet file viewer

Deephaven Parquet Viewer

A Parquet viewer powered by deephaven-core. Supports opening and browsing through large (or small) parquet files.

Helper script

Usage: deephaven-parquet-viewer.sh [parquet-file] ([port])

Docker container

docker run \
    --rm \
    --name deephaven-parquet-viewer \
    --mount type=bind,source=/path/to/file.parquet,target=/file.parquet,readonly \
    -p 10000:10000 \
    ghcr.io/devinrsmith/deephaven-parquet-viewer:latest

Browser

• Table iframe: http://localhost:10000/iframe/table/?name=parquet_table
• IDE: http://localhost:10000/ide/

Requirements

Linux, Mac OS (Intel or M1), and Windows WSL should all be supported. Docker is currently a requirement for running the helper script and via a Docker container, but advanced users can get the application running natively if desired.

deephaven-parquet-viewer's Issues

Improve error message when container fails to start

The script assumes that the container will start successfully. When that's not the case, the health check will keep looping every second.

It might be necessary to remove the --rm and relay any container error messages to the user (and then, do our own docker rm in the cleanup trap).

No such object: deephaven-parquet-viewer

Just cloned this on my Mac and started Docker. Then I ran it like this:

./deephaven-parquet-viewer.sh some-file.parquet

It installed the docker image and then it started showing errors like this:

Starting Deephaven Parquet Viewer......Error: No such object: deephaven-parquet-viewer
.Error: No such object: deephaven-parquet-viewer
.Error: No such object: deephaven-parquet-viewer
.Error: No such object: deephaven-parquet-viewer
.Error: No such object: deephaven-parquet-viewer

The "No such object" errors kept coming, about once a second.

Please help, thanks.

token

How to decide token at first?

Is there a way to save in JSON or CSV?

I was wondering if the viewer can save in JSON or CSV.