serset.hlp

是一个经济学管理应用软件 很难找的 但是经济学学生又必须用到

{smcl}
{* 10feb2005}{...}
{cmd:help serset}
{hline}

{title:Title}

{p2colset 5 19 21 2}{...}
{p2col :{hi:[P] serset} {hline 2}}Creation and manipulation of sersets{p_end}
{p2colreset}{...}


{title:Syntax}

{p 8 22 2}
{cmd:serset}
{cmdab:cr:eate}
{varlist}
{ifin}
[{cmd:,}
{cmdab:omitanym:iss}
{cmdab:omitallm:iss}
{cmdab:omitdupm:iss}
{cmdab:omitn:othing}
{opth sort(varlist)}]

{p 8 22 2}
{cmd:serset}
{cmd:create_xmedians}
{it:svn_y}
{it:svn_x}
[{it:svn_w}]
[{cmd:,}
{cmd:bands(}{it:#}{cmd:)}
{cmd:xmin(}{it:#}{cmd:)}
{cmd:xmax(}{it:#}{cmd:)}
{cmd:logx}
{cmd:logy}]

{p 8 22 2}
{cmd:serset}
{cmd:create_cspline}
{it:svn_y}
{it:svn_x}
[{cmd:,}
{cmd:n(}{it:#}{cmd:)}]

{p 8 22 2}
{cmd:serset}
[{cmd:set}]
{it:#_s}

{p 8 22 2}
{cmd:serset}
{cmd:sort}
[{it:svn} [{it:svn} [...]]]

{p 8 22 2}
{cmd:serset}
{cmdab:su:mmarize}
{it:svn}
[{cmd:,}
{cmdab:d:etail}
]

{p 8 22 2}
{cmd:serset}
{cmd:use}
[{cmd:,}
{cmd:clear}
]

{p 8 22 2}
{cmd:serset}

{p 8 22 2}
{cmd:serset}
{cmd:reset_id}
{it:#_s}

{p 8 22 2}
{cmd:serset}
{cmd:drop}
[{numlist} | {cmd:_all}]

{p 8 22 2}
{cmd:serset}
{cmd:clear}

{p 8 22 2}
{cmd:serset}
{cmd:dir}


{pstd}
In addition, the {helpb file} command is extended to allow

{p 8 22 2}
{cmd:file}
{cmd:sersetwrite}
{it:handle}

{p 8 22 2}
{cmd:file}
{cmd:sersetread}
{it:handle}


{pstd}
In addition, the following {help macro:extended macro functions} are available

	Extended function             Returns from the {it:current serset}
	{hline 61}
	{cmd:: serset id}                   ID
	{cmd:: serset k}                    number of variables
	{cmd:: serset N}                    number of observations
	{cmd:: serset varnum} {it:svn}           {it:svnum} of {it:svn}
	{cmd:: serset type} {it:svn}             storage type of {it:svn}
	{cmd:: serset format} {it:svn}           display format of {it:svn}
	{cmd:: serset varnames}             list of {it:svns}
	{cmd:: serset min} {it:svn}              minimum of {it:svn}
	{cmd:: serset max} {it:svn}              maximum of {it:svn}
	{hline 61}
	Extended macro functions have the syntax
		{cmd:local} {it:macname} {cmd::} ...
	The {it:current serset} is the most recently created or the most
		recently set by the {cmd:serset set} command.


{pstd}
In the above syntax diagrams,

{phang2}
{it:#_s} refers to a serset number, 0 {ul:<} {it:#} {ul:<} 1,999.

{phang2}
{it:varlist} refers to the usual Stata {help varlist}, that is, a list of
variables that appear in the current dataset, not the current serset.

{phang2}
{it:svn} refers to a variable in a serset.  The variable may be referred to
either by its name (e.g., {cmd:mpg} or {cmd:l.gnp}) or its number (e.g., 1 or
5); which is used makes no difference.

{phang2}
{it:svnum} refers to a variable number in a serset.


{title:Description}

{pstd}
{cmd:serset} creates and manipulates sersets.

{pstd}
{cmd:file} {cmd:sersetwrite} writes and {cmd:file} {cmd:sersetread} reads
sersets into files.

{pstd}
The extended macro function {cmd::serset} reports information about the
current serset.


{title:Options for serset create}

{phang}
{cmd:omitanymiss}, {cmd:omitallmiss}, {cmd:omitdupmiss}, and {cmd:omitnothing}
    specify how observations with missing values are to be treated.

{pmore}
    {cmd:omitanymiss} is the default.  Observations in which any of the
    numeric variables contain missing are omitted from the serset being
    created.

{pmore}
    {cmd:omitallmiss} specifies that only observations in which all the
    numeric variables contain missing are to be omitted.

{pmore}
    {cmd:omitdupmiss} specifies that only duplicate observations in which all
    the numeric variables contain missing are omitted.  Note that the
    observations omitted will be a function of the sort order of the original
    data.

{pmore}
    {cmd:omitnothing} specifies that no observations be omitted
    (other than those excluded by the {cmd:if} {it:exp} and {cmd:in}
    {it:range}).

{phang}
{cmd:sort(}{it:varlist}{cmd:)} specifies that the serset being created is to
   be sorted by the specified variables.  The result is no different from,
   after serset creation, using the {cmd:serset sort} command, but total
   execution time is a little faster.  The sort order of the data in memory is
   unaffected by this option.


{title:Options for serset create_xmedians}

{phang}
{cmd:bands(}{it:#}{cmd:)} specifies the number of divisions along the x-scale
    in which cross medians are to be calculated; the default is
    {cmd:bands(200)}.  {cmd:bands()} may be specified to be between 3 and 200.

{pmore}
    Let {it:m} and {it:M} specify the minimum and maximum value of x.
    If the scale is divided into {it:n} bands (i.e., {cmd:bands(}{it:n}{cmd:)}
    is specified), the first band is
    {it:m} to
    {it:m}+({it:M}-{it:m})/{it:n}, the second
    {it:m}+({it:M}-{it:m})/{it:n} to
    {it:m}+2*({it:M}-{it:m})/{it:n}, ..., and the nth
    {it:m}+({it:n}-1)*({it:M}-{it:m})/{it:n} to
    {it:m}+{it:n}*({it:M}-{it:m})/{it:n} = {it:m}+{it:M}-{it:m} = {it:M}.

{phang}
{cmd:xmin(}{it:#}{cmd:)} and {cmd:xmax(}{it:#}{cmd:)} specify the minimum and
    maximum values of the x variable to be used in the bands
    calculation{hline 2}{it:m} and {it:M} in the formulas above.  The actual
    minimum and maximum are used if these options are not specified.  In
    addition, if {cmd:xmin()} is specified with a number that is greater than
    the actual minimum, the actual minimum is used, and if {cmd:xmax()} is
    specified with a number that is less than the actual maximum, the actual
    maximum is used.

{phang}
{cmd:logx} and {cmd:logy} specify that cross medians are to be created using a
    "log" scale.  The exponential of the median of the log of the values is
    calculated in each band.


{title:Option for serset create_cspline}

{phang}
{cmd:n(}{it:#}{cmd:)} specifies the number of points to be evaluated between
    each pair of x values, which are treated as the knots.  The default is
    {cmd:n(5)}, and {cmd:n()} may be between 1 and 300.


{title:Option for serset summarize}

{phang}
{cmd:detail} specifies additional statistics, including skewness,
    kurtosis, the four smallest and four largest values, and various
    percentiles.  This option is identical to the {cmd:detail} option of
    {helpb summarize}.


{title:Option for serset use}

{phang}
{cmd:clear} permits the serset to be loaded, even if there is a dataset already
    in memory and even if that dataset has changed since it was last saved.


{title:Remarks}

{pstd}
Sersets are used in implementing Stata's graphics capabilities.  When you make
a graph, the data for the graph are extracted into a serset and then, at the
lowest levels of Stata's graphics implementation, are graphed from there.

{pstd}
Sersets are like datasets:  they contain observations on one or more
variables.  Each serset is assigned a number, and, in your program, you use
that number when referring to a serset.  Thus multiple sersets can reside
simultaneously in memory.  (Sersets are, in fact, stored in a combination of
memory and temporary disk files, so accessing their contents is slower than
accessing the data in memory.  Sersets, however, are fast enough to keep up
with graphics operations.)


{title:serset create}

{pstd}
{cmd:serset create} creates a new serset from the data in memory.  For
instance,

	{cmd:. serset create mpg weight}

{pstd}
creates a new serset containing variables {cmd:mpg} and {cmd:weight}.  When
using the serset subsequently, you can refer to these variables by their
names, {cmd:mpg} and {cmd:weight}, or by their numbers, {cmd:1} and {cmd:2}.

{pstd}
{cmd:serset create} also returns in {cmd:r()}:

	{cmd:r(N)}    the number of observations placed into the serset
	{cmd:r(k)}    the number of variables placed into the serset
	{cmd:r(id)}   the number assigned to the serset

{pstd}
{cmd:r(N)} and {cmd:r(k)} are just for your information; by far the most
important returned result is {cmd:r(id)}.  You will need to use this number
in subsequent commands to refer to this serset.

{pstd}
In addition, {cmd:serset create} sets the current serset to the one just
created.  Commands that use sersets always use the current serset.  If, in