KitForma field notes

Preserve numeric types when exporting HDF5 data

Separate exact binary arrays from convenient text tables and previews.

Use NPY for primitive numeric arrays

NPY stores a numeric dtype, shape, storage order and binary values. This workbench writes primitive integer and float datasets in little-endian order after reading their values through the HDF5 engine. The original byte order is recorded in the report, so a big-endian source does not become an undocumented interpretation change.

The output retains selected axes, including length-one axes. Integer widths and signedness remain explicit. A 64-bit identifier is not converted through a JavaScript floating-point number before being written. The independent fixture checks include integers beyond the safe JavaScript integer range and a big-endian compressed numeric cube.

Treat CSV as an import decision

CSV is useful for inspecting coordinates and values but carries no universal numeric type system. Exact 64-bit integer values are written as decimal strings. A spreadsheet may still infer a floating-point number and round it when opening the file. Import such columns as text or use NPY/JSON when exact type retention matters.

Strings that resemble formulas are neutralized in CSV by the shared safe-export rule. The unmodified values remain in JSON. Flat compound datasets have one output field per supported member, while nested or variable-length compound structures are metadata-only. The report explains which source features were not represented in the destination.

Do not measure from a normalized PNG

The optional PNG maps the finite minimum and maximum of a small slice to grayscale. It is meant for visual inspection and can help reveal an orientation or gross pattern. Nonfinite values are transparent, and a constant finite slice uses middle gray. The output report states the normalization range and dimensions.

A normalized image is not quantitative scientific data, and two independently normalized images cannot be compared by brightness alone. Keep NPY or JSON for calculations. The workbench excludes 64-bit integer PNG previews to avoid suggesting exact visual resolution for values that cannot all be distinguished in an 8-bit grayscale image.

The evidence behind the check

Sources & assumptions

Formats, algorithms and supported features are documented below. Check the stated scope and inspect your output before using it.

  • h5wasm — NIST ↗

    Unmodified h5wasm 0.10.3 provides HDF5 C-library access. NIST is acknowledged as the source; no endorsement is claimed.

  • HDF5 documentation ↗

    Groups, datatypes, attributes, links, filters and datasets are distinct HDF5 concepts.

  • NumPy NPY format ↗

    NPY records array shape, order, dtype and binary values.

  • Engine notices ↗

    Full NIST/HDF5 and dependency notices are shipped with the local engine.

Reviewed 2026-10-19. Read our methodology or report an issue.

Essential local storage remembers this choice for up to 180 days. Your choice applies to this subdomain. Global Privacy Control and Do Not Track override optional measurement.

Advertising has a separate choice. Use Advertising preferences in the footer to allow or reject the optional guide-page banner.