filestat.tex

\documentclass{scrartcl}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{amsmath}
\usepackage{amssymb}
\usepackage{booktabs}
\usepackage{mathtools}
\usepackage{multirow}
\usepackage{pgfplots}
\usepackage{pgfplotstable}
\usepackage{longtable}
\usepackage{threeparttable}
\usepackage{ifthen}
\usepackage{url}
\usepackage{hyperref}
\pgfplotsset{compat=1.9}

\title{\vspace{-5ex}Test data statistics}
\author{}
\date{}
\begin{document}
\thispagestyle{empty}
\maketitle
\thispagestyle{empty}

\vspace{-10ex}
\centering

%%%% table setup
\pgfplotstableread{filestat.dat}\filestat
\pgfplotstablecreatecol[
	create col/assign/.code={%
		\getthisrow{category}\entry
		\getprevrow{category}\preventry
		\ifthenelse{\equal{\entry}{\preventry}}{
			\edef\corpus{}
		}{
			\edef\corpus{\entry}
		}
		\pgfkeyslet{/pgfplots/table/create col/next content}\corpus
	}]{corpus}\filestat

\begin{threeparttable}
\newcounter{corpuscount}
\setcounter{corpuscount}{1}
\renewcommand{\arraystretch}{1.15}
\setlength{\tabcolsep}{1em}
\scriptsize
\pgfplotstabletypeset[
	every head row/.style={
		before row={\toprule},
		after row={\midrule}},
	every last row/.style={
		after row={\bottomrule}},
	columns={corpus,file,size,sigma,newlines},
	columns/file/.style={string type,column type=l,column name={File}},
	columns/sigma/.style={fixed,column type=r,column name={Alphabet size}},
	columns/newlines/.style={fixed,column type=r,column name={Lines}},
	columns/size/.style={fixed,fixed zerofill,precision=3,preproc/expr={##1/(1024*1024)},column type=r,column name={Size in MB}},
	columns/corpus/.style={string type,column type=l,column name={Text corpus},
		postproc cell content/.append style={
			/pgfplots/table/@cell content/.add={\,\ifthenelse{\equal{##1}{}}{}{%
				\tnote{\thecorpuscount}\addtocounter{corpuscount}{1}%
				}
			}
		}
	}
	]\filestat
%%%% table notes with URLs
\setcounter{corpuscount}{1}
\pgfplotstabletypeset[
	begin table={\begin{tablenotes}}, end table={\end{tablenotes}},
	before row={\item[\thecorpuscount]  \addtocounter{corpuscount}{1}},
	after row={},
	every head row/.style={before row={\iffalse},after row={\fi}},
	typeset cell/.style={
		/pgfplots/table/@cell content={\href{#1}{#1}}
	},
	skip coltypes,
	TeX comment=,
	columns={source},
	columns/source/.style={string type},
	unique={category}]\filestat
\end{threeparttable}

\end{document}