r/man/Scanner.Rd

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/dataset.R
\name{Scanner}
\alias{Scanner}
\alias{ScannerBuilder}
\title{Scan the contents of a dataset}
\description{
A \code{Scanner} iterates over a \link{Dataset}'s fragments and returns data
according to given row filtering and column projection. A \code{ScannerBuilder}
can help create one.
}
\section{Factory}{

\code{Scanner$create()} wraps the \code{ScannerBuilder} interface to make a \code{Scanner}.
It takes the following arguments:
\itemize{
\item \code{dataset}: A \code{Dataset} or \code{arrow_dplyr_query} object, as returned by the
\code{dplyr} methods on \code{Dataset}.
\item \code{projection}: A character vector of column names to select
\item \code{filter}: A \code{Expression} to filter the scanned rows by, or \code{TRUE} (default)
to keep all rows.
\item \code{use_threads}: logical: should scanning use multithreading? Default \code{TRUE}
\item \code{...}: Additional arguments, currently ignored
}
}

\section{Methods}{

\code{ScannerBuilder} has the following methods:
\itemize{
\item \verb{$Project(cols)}: Indicate that the scan should only return columns given
by \code{cols}, a character vector of column names
\item \verb{$Filter(expr)}: Filter rows by an \link{Expression}.
\item \verb{$UseThreads(threads)}: logical: should the scan use multithreading?
The method's default input is \code{TRUE}, but you must call the method to enable
multithreading because the scanner default is \code{FALSE}.
\item \verb{$BatchSize(batch_size)}: integer: Maximum row count of scanned record
batches, default is 32K. If scanned record batches are overflowing memory
then this method can be called to reduce their size.
\item \verb{$schema}: Active binding, returns the \link{Schema} of the Dataset
\item \verb{$Finish()}: Returns a \code{Scanner}
}

\code{Scanner} currently has a single method, \verb{$ToTable()}, which evaluates the
query and returns an Arrow \link{Table}.
}