{smcl}
{* *! version 1.0  25Feb2009}{...}
{hline}
help for {hi:ifwins} {right:manual:  {hi:[R] none}}
{right:dialog:  {hi: none}    }
{hline}


{title:subset dataset by if exp before subsetting by in range}

{p 8 17 2}{cmd:ifwins}
{ifin}
{cmd::} {it:stata_cmd}{p_end}

        
{title:Description}

{p 4 4 2}{cmd:ifwins} is a prefix command that runs most any Stata command that does not
modify the dataset in memory (e.g. {help generate:generate}, {help replace:replace}, etc.).  
{cmd:ifwins} will have {help if:if} subset the dataset before {help in:in} subsets the dataset.  
This is the opposite of what happens when both {help if:if} and {help in:in} are used in the 
same Stata command. For example, the following code will first subset the dataset to the first 
10 observations and then subset the dataset to the specified condition:{p_end}

{p 4 8 2}{cmd:. sysuse auto}{p_end}
{p 4 8 2}{cmd:. list  if foreign == 1 in 1/10}{p_end}

{p 4 4 2}Since the auto.dta dataset is sorted by the variable {cmd:foreign}, the above code will 
not list any observations because in the first 10 observations {cmd:foreign == 0} .  So, {help if:if}
looses and {help in:in} wins when "battling" over which one subsets the dataset before the other one does.{p_end}

{p 4 4 2}If you want to run a Stata command on a certain number of observations when a certain condition exists, 
you would have to:{p_end}

{p 4 8 2}{cmd:. preserve}{p_end}
{p 4 8 2}{cmd:. keep if foreign == 1 }{p_end}
{p 4 8 2}{cmd:. list in 1/10 make turn weight }{p_end}
{p 4 8 2}{cmd:. restore}{p_end}

{p 4 4 2}
or use {cmd:ifwins} as a prefix to the desired Stata command:{p_end}

{p 4 8 2}{cmd:. ifwins  if foreign == 1 in 1/10 : list make turn weight }{p_end}

{p 4 4 2}The above will {help list:list} up to the first 10 observations of when the variable {cmd:foreign}
is equal to 1 (one).  So now {help if:if} wins!...but {help in:in} is still helpful.{p_end}

{title:Remarks}

{p 4 4 2}{help if:if} and {help in:in} are not allowed in the Stata command {cmd:ifwins} is specified to run.{p_end}
 
{p 4 4 2}Since {cmd:ifwins} (when not running the {help list:list}, {help browse:browse}, {help edit:edit} commands) 
temporarily subsets the dataset before 
running the submitted Stata command the observations numbers will not be the same as in the original dataset.  
So, the {help _variables:system variables} {help _n:_n} and {help _N:_N} as well as the {help creturn:c(N)} macro 
variable are reset to the subsetted dataset before the specified Stata command is run by {cmd:ifwins}. Also, 
{cmd:ifwins} will never return the error message "Obs. nos. out of range {search r(198):r(198)}" if the 
range specified is at least within the number of observations in the dataset.  {cmd:ifwins}
corrects ranges specified by {help in:in} to at least 1 observation of the specified condition since
it is hard to know how many observations will be available after the {help if:if} condition subsets
the dataset.{p_end}

{p 4 4 2}The {help list:list}, {help browse:browse}, {help edit:edit} commands are the only exceptions to the note above.
{p_end}

{title:Examples}

{p 4 8 2}{cmd:. ifwins  if mpg < 25  in -10/L : list make turn weight }{p_end}

{p 4 8 2}{cmd:. ifwins  if income > 1000000 in 1/200 : summarize x1 x2 x3 x4 year89 year90 }{p_end}


{title:Author}

{p 4 4 2}
Dan Blanchette {break}
The Carolina Population Center {break}
University of North Carolina - Chapel Hill, USA {break}
dan_blanchette@unc.edu{p_end}


{title:Also see}

{p 4 13 2}On-line: {help if:if} and {help in:in}{p_end}