Skip to content

HDF5.to_csv

to_csv(h5file, outdir, sep=" ")

Converts the datasets in an HDF5 file into CSV files inside outdir. The file is loaded into an HDF5Dict and its tree is walked group by group; how a dataset is written depends on its _kind:

  • linear datasets — all linear datasets found at the same group level are treated as columns and concatenated side by side into a single CSV named after that group (<group>.csv; the root level falls back to out.csv). Each column header is the dataset's _name attribute, or its key if _name is absent.
  • compound datasets — each one is written to its own CSV file, since a structured array already holds multiple named columns. The file is named after the dataset's _name attribute, or its key (<name>.csv).

Sub-groups are traversed recursively, so a nested HDF5 hierarchy produces one CSV per group of linear columns plus one CSV per compound dataset. _attrs metadata is not exported.

Parameters

h5file : str or Path
Source HDF5 file.
outdir : str or Path
Destination directory, created if it does not exist.
sep : str, optional
Column separator. Default " ".

Example

Given an HDF5 file rates.hdf5 with this structure:

/
└── rates/                  (group)
    ├── T            linear   _name="T"
    ├── k1           linear   _name="k1"
    ├── k2           linear   _name="k2"
    └── output_units compound _name="output_units"  (col1, col2, col3)

from jaff.drivers import HDF5

HDF5().to_csv("rates.hdf5", "out/", sep=",")

produces two files in out/:

  • rates.csv — the three linear datasets merged as columns:
T,k1,k2
100.0,1.2e-10,3.4e-11
200.0,2.3e-10,4.5e-11
  • output_units.csv — the compound dataset, one column per field:
col1,col2,col3
1.5,3,cm3/s
2.5,4,s-1