Skip to contents

Write (updated) values back into existing HDF5 datasets (robust)

Source: R/hdf5_read-write.R
h5_write_values.Rd

Writes scalars, vectors, matrices/arrays, and 2-column data frames/tibbles (treated as time series) into existing HDF5 datasets. If the dataset reports SCALAR incorrectly, the function can infer target dimensions from the supplied value and resize accordingly.

Usage

h5_write_values(
  h5,
  values,
  resize = TRUE,
  strict = TRUE,
  prefer_rows = NA,
  ts_cols = c("time", "value"),
  scalar_strategy = c("error", "first", "collapse"),
  collapse_sep = ";",
  verbose = FALSE
)

Arguments

h5

An open hdf5r::H5File (e.g., mode = "r+").

values

Named list: names are absolute dataset paths, values are R objects to write.

resize

Logical. If TRUE, resize datasets via set_extent() when shapes differ.

strict

Logical. If TRUE, stop on first error; otherwise warn and skip.

prefer_rows

Logical(1) or NA. For 2-column time series: NA keeps dataset orientation (2xN if first dim == 2), TRUE forces 2xN, FALSE forces Nx2.

ts_cols

Character(2). Column names to pull from time-series data frames (default c("time","value")).

scalar_strategy

One of "error", "first", "collapse". Controls how non-length-1 values are handled for true SCALAR datasets.

collapse_sep

Character. Separator used when scalar_strategy = "collapse".

verbose

Logical. If TRUE, prints per-path dimension info.

Value

Invisibly returns the character vector of written dataset paths.

Examples

if (FALSE) { # \dontrun{
vals <- h5_read_values(h5)
vals[["/Parameters/OutputPath"]] <- "C:/temp/out.h5"
h5_write_values(h5, vals, resize = TRUE, scalar_strategy = "first", verbose = TRUE)
} # }