What HDF5 files are supported?

HDF5 files with at least one structured (table-like) dataset in the root group — the first one found is converted. Plain arrays without a compound/structured layout have no table-like dataset to convert and will fail with a clear diagnostic. Files are limited to 40 MB.

Which columns and values come through?

The JSON report includes each column name, its datatype, any unit annotation, and all row values of the first structured dataset, plus the dataset's metadata.

Is my data shared or published?

No. The file is processed when you press Convert file and the JSON report is returned to you; nothing is indexed or published.

Convert an HDF5 (.h5) file to JSON online

Need to peek inside an HDF5 (.h5) file without installing h5py or pulling up h5dump? This page reads the structured datasets in an HDF5 file with Astropy and returns them as JSON you can search, copy, or feed into your own scripts.

How to convert

Upload an .h5 (or .hdf5, .hd5, …) file of up to 40 MB.
Press Convert file.
Review the JSON report: dataset path, row/column counts, per-column datatype and unit, all values, and dataset metadata.

What you get

A compact JSON object describing the first structured dataset in the file:

format — always hdf5
path — dataset name inside the file
n_rows / n_columns — shape of the table
columns — for each column: name, dtype, unit, and the full value list
meta — any stored dataset attributes

When this helps

Inspecting HDF5 exports from instruments, simulations, or data pipelines
Converting small scientific tables for spreadsheets, notebooks, or plain JSON tooling
Sanity-checking what a .h5 file actually contains before writing code against it

Works only on genuine HDF5 files with a structured dataset — malformed or non-HDF5 uploads get an explicit error instead of garbage output.

Operation details

What happens: Astropy opens the HDF5 container, finds the first structured (table-like) dataset, and serializes column names, dtypes, units, attributes, and all row values to JSON.

Engine: Astropy HDF5 reader.

Inputs: .h5, .hdf5, or .hd5 up to 40 MB.

Output: JSON describing one dataset path, row/column counts, per-column values, and dataset metadata. Plain numeric arrays without a compound layout are rejected with a diagnostic.

Lossy? Read-only inspection — no bytes in the file are modified.

Common failures: Files with no structured dataset, corrupt HDF5, or non-HDF5 uploads.

Use when: You need to inspect instrument or simulation tables before writing h5py code.