{smcl}
{* 24 April 2006}{...}
{hline}
help for {hi:tabexport}
{hline}

{title:Export tables of summary statistics to text files}

{p 8 17 2}
{cmd:tabexport} 
{it:varlist} 
[{cmd:if} {it:exp}] 
[{cmd:in} {it:range}] 
[{it:weight}] 
{cmd:using} {it:filename}
{cmd:,}
{cmdab:s:tatistics(}{it:statistics_list}{cmd:)} 
[
{cmdab:l:ist(}{it:list_options}{cmd:)} 
{cmdab:f:ormat(}{it:format(s)}{cmd:)} 
{cmd:by(}{it:varlist}{cmd:)}
{cmd:cw} 
{cmd:total} 
{cmd:noreshape}
{cmdab:var:iableorder}
{cmd:outfile}
{it:out_options}
]  

{p 4 4 2}
Any weights are allowed, except as indicated by help on {help collapse}.


{title:Description}

{p 4 4 2}
{cmd:tabexport} exports as tables to text files various arrays of summary
statistics which could be created by using {help collapse}. By default 
{help outsheet} is used, but optionally {help outfile} may be specified. 
The results are also {cmd:list}ed to provide a check on what is produced. 


{title:Options} 

{p 4 8 2}{cmd:statistics()} is a required option. One or more 
statistic names as specified in the help for {help collapse} should be used. 
Note that each statistic will be calculated for each variable: 
{cmd:collapse} itself is not so restrictive. 

{p 4 8 2}{cmd:format()} specifies one or more numeric formats controlling
the detail of each statistic produced. The sequence of formats is
that of the sequence of statistics. Thus with 
{cmd:s(count mean) f(%9.0f %9.1f)} 
counts will be shown with format %9.0f and means with format %9.1f. If fewer
formats are specified than there are statistics, the last-specified format
will be repeated for all remaining statistics. Thus with 
{cmd:s(count mean sd) f(%9.0f %9.1f)} 
counts will be shown with format %9.0f and means and sds with format %9.1f. 
See help on {help format}. 

{p 4 8 2} 
{cmd:list()} specifies options of {help list} to tune the output listing. 

{p 4 8 2}
{cmd:by(}{it:byvarlist}{cmd:)} and {cmd:cw} are the 
corresponding options of {help collapse}.

{p 4 8 2}
{cmd:total} applies whenever {cmd:by()} is also specified and specifies that
results should be appended for all of the observations specified. Such results
will be listed separately, as a reflection of the production process, but
they will be combined in the resulting file. 

{p 4 8 2}
{cmd:noreshape} specifies that files produced should have a wide data
structure, with many columns (variables).  The default is that they should be
{cmd:reshape}d to have a long data structure, with fewer columns.  

{p 4 8 2}
{cmd:variableorder} may be specified with {cmd:noreshape}. The default
of {cmd:noreshape} is that of {help collapse}: the variable order is by 
statistic, then by original variable. {cmd:variableorder} reverses
this: the variable order is by original variable, then by statistic. 

{p 4 8 2} 
{cmd:outfile} specifies that {help outfile} should be used to 
export files. The default is to use {help outsheet}. 
 
{p 4 8 2}{it:out_options} are options of {cmd:outsheet} 
or (if the {cmd:outfile} option is specified) of {cmd:outfile}. 
By default, {cmd:noquote} is specified. 


{title:Examples}

{p 4 8 2}{cmd:. sysuse auto, clear}{p_end}
{p 4 8 2}{cmd:. tabexport turn trunk length using results.txt, s(mean sd) by(foreign)}{p_end}
{p 4 8 2}{cmd:. tabexport turn trunk length using results.txt, s(mean sd) by(foreign) total replace}{p_end}
{p 4 8 2}{cmd:. tabexport turn trunk length using results.txt, s(mean sd) by(foreign) total replace format(%9.1f)}{p_end}
{p 4 8 2}{cmd:. type results.txt}


{title:Author} 

{p 4 4 2}Nicholas J. Cox, Durham University, U.K.{break} 
         n.j.cox@durham.ac.uk


{title:Acknowledgements} 

{p 4 4 2}Raymond Guiteras suggested this problem. Kit Baum provided some 
helpful field testing, identified an ordering problem and suggested the 
{cmd:format()} option. Fred Wolfe suggested adding a handle for {cmd:list} 
and identified another ordering problem. 


{title:Also see}

{p 4 13 2}On-line: help for {help tabstat}