% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/de_serialize.R
\name{de_serialize}
\alias{de_serialize}
\alias{binarize}
\alias{binarizeTransform}
\alias{getDataInFile}
\alias{binarize}
\alias{binarizeTransform}
\alias{getDataInFile}
\title{(De)Serialization of a [big]matrix or data stream}
\usage{
binarize(data_ascii, data_bin_file, nb_per_chunk, sep = ",", nbytes = 4,
  endian = .Platform$endian)

binarizeTransform(getData, transform, data_bin_file, nb_per_chunk, nbytes = 4,
  endian = .Platform$endian)

getDataInFile(indices, data_bin_file, nbytes = 4, endian = .Platform$endian)
}
\arguments{
\item{data_ascii}{Either a matrix or CSV file, with items in rows}

\item{data_bin_file}{Name of binary file on output (\code{binarize})
or intput (\code{getDataInFile})}

\item{nb_per_chunk}{Number of lines to process in one batch}

\item{sep}{Separator in CSV input file (if any provided)}

\item{nbytes}{Number of bytes to serialize a floating-point number; 4 or 8}

\item{endian}{Endianness to use for (de)serialization. Use "little" or "big" for portability}

\item{getData}{Function to retrieve data chunks}

\item{transform}{Transformation function to apply on data chunks}

\item{indices}{Indices of the lines to retrieve}
}
\value{
For \code{getDataInFile()}, the matrix with rows corresponding to the
  requested indices. \code{binarizeTransform} returns the number of processed lines.
  \code{binarize} is designed to serialize in several calls, thus returns nothing.
}
\description{
\code{binarize()} serializes a matrix or CSV file with minimal overhead,
  into a binary file. \code{getDataInFile()} achieves the inverse task: she retrieves
  (ASCII) data rows from indices in the binary file. Finally,
  \code{binarizeTransform()} serialize transformations of all data chunks; to use it,
  a data-retrieval function must be provided, thus \code{binarize} will most likely be
  used first (and then a function defined to seek in generated binary file)
}