{smcl}
{* *! version 0.1  14sep2010}{...}
{cmd:help readreplace}
{hline}

{title:Title}

    {hi:readreplace} -- Make many replacements to a dataset from a .csv file

{title:Syntax}

{p 8 17 2}
{cmdab:readreplace}
{cmd: using} {it: filename}
{cmd:, id}({varname}) 
[{cmd:display}]

{synoptset 20 tabbed}{...}
{synopthdr}
{synoptline}
{synopt:{opt id(varname)}}Required. Name of unique identifier{p_end}
{synopt:{opt di:splay}}show detail on what readreplace is doing{p_end}
{synoptline}
{p2colreset}{...}
{p 4 6 2}

{title:Description}

{pstd}
{cmd:readreplace} makes replacements using a very specifically-formated .csv file. 
	It is intended for use with the output from {help cfout}.
	
{title:Remarks}

{pstd}
{cmd: readreplace} is intended to be used as part of the data entry process
	when data is entered two times for accuracy.  
	After the second entry, the datasets need to be reconciled.  
	{help cfout} will compare the first and second entries and 
	generate a list of discrepancies in a format that is useful 
	for the data entry teams. 
	The data entry operators can then simply type the correct value 
	in a new column of the cfout results. 
	To make the changes in your dataset, 
	load your data then run readreplace	using the new .csv file.
	
{pstd}
{cmd: readreplace} requires the using file to be a .csv file with the format
	
{center:{it: id value, question name , correct value } }

{pstd}
	It then runs the replaces
	
{center: replace {it:question name} = {it:correct value} if {it:id varname} == {it:id value} }

{pstd}
	for each line in the csv file, 
	with allowances for whether the variables/ids are string or numeric.
	The display option is useful for debugging. 
	The first row is assumed to be a header row, and no replacements are made to data in the first row.
	
{title:Examples}

{phang}readreplace using "corrected values.csv", id(uniqueid)

{title:Also see}

{psee}
Online:  {help cf}, {help cfout}
{p_end}