addplot_option.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}
{* 04mar2005}{...}
{cmd:help addplot_option}
{hline}
{* index addplot() tt option}{...}

{title:Title}

{p2colset 5 27 29 2}{...}
{p2col :{hi:[G] {it:addplot_option}} {hline 2}}Option for adding additional twoway plots to command{p_end}
{p2colreset}{...}


{title:Syntax}

{p 8 16 2}
{it:command}
...
[{cmd:,}
...
{cmd:addplot(}{it:plot} ... [{cmd:||} {it:plot} ... [...]]{cmd:)}
...]

{phang}
where {it:plot} may be
any subcommand of {cmd:graph} {cmd:twoway},
such as
{cmd:scatter},
{cmd:line}, {cmd:histogram}, etc.


{title:Description}

{pstd}
Some commands that draw graphs (but do not start with the word {cmd:graph})
are documented in the other reference manuals.  Many of those commands allow
the {cmd:addplot()} option.  This option allows them to overlay their results
on top of {cmd:graph} {cmd:twoway} plots.


{title:Option}

{phang}
{cmd:addplot(}{it:plot}{cmd:)}
    specifies the rest of the {cmd:graph} {cmd:twoway} subcommands to be
    added to the {cmd:graph} {cmd:twoway} command issued by {it:command}.


{title:Remarks}

{pstd}
Remarks are presented under the headings

	{help addplot_option##remarks1:Commands that allow the addplot() option}
	{help addplot_option##remarks2:Advantage of graph twoway commands}
	{help addplot_option##remarks3:Advantages of graphic commands implemented outside of graph twoway}
	{help addplot_option##remarks4:Use of the addplot() option}


{marker remarks1}{...}
{title:Commands that allow the addplot() option}

{pstd}
{cmd:graph} commands never allow the {cmd:addplot()} option.  The 
{cmd:addplot()} option is allowed by commands outside of {cmd:graph} that are
implemented in terms of {cmd:graph} {cmd:twoway}.

{pstd}
For instance, the {cmd:histogram} command{hline 2}see
{helpb histogram}{hline 2}allows {cmd:addplot()}.  {cmd:graph} {cmd:twoway}
{cmd:histogram}{hline 2}see {helpb twoway histogram}{hline 2}does not.


{marker remarks2}{...}
{title:Advantage of graph twoway commands}

{pstd}
The advantage of {cmd:graph} {cmd:twoway} commands is that they can be
overlaid, one on top of the other.  For instance, you can type

{phang2}
	{cmd:. graph twoway scatter} {it:yvar} {it:xvar} {cmd:||}
		{cmd:lfit} {it:yvar} {it:xvar}

{pstd}
and the separate graphs produced, {cmd:scatter} and {cmd:lfit}, are
combined into one.
The variables to which each refers need not even be the same:

{phang2}
	{cmd:. graph twoway scatter} {it:yvar} {it:xvar} {cmd:||}
		{cmd:lfit} {it:y2var} {it:x2var}


{marker remarks3}{...}
{title:Advantages of graphic commands implemented outside of graph twoway}

{pstd}
Graphic commands implemented outside of {cmd:graph} {cmd:twoway} can
have simpler syntax.  For instance, the {cmd:histogram} command
has an option, {cmd:normal}, that will overlay a normal curve on top of
the histogram:

	{cmd:. histogram} {it:myvar}{cmd:, normal}

{pstd}
That is easier than typing

	{cmd:. summarize} {it:myvar}
	{cmd:. graph twoway histogram} {it:myvar} {cmd:||}
	    {cmd:function} {cmd:normden(x,`r(mean)',`r(sd)'),} {cmd:range(}{it:myvar}{cmd:)}

{pstd}
which is the {cmd:graph} {cmd:twoway} way of producing the same thing.

{pstd}
Thus the trade-off between {cmd:graph} and non-{cmd:graph} commands is one of
greater flexibility versus easier use.


{marker remarks4}{...}
{title:Use of the addplot() option}

{pstd}
The {cmd:addplot()} option attempts to give back flexibility to non-{cmd:graph}
graphic commands.  Such commands are, in fact, implemented in terms of
{cmd:graph} {cmd:twoway}.  For instance, when you type

	{cmd:. histogram} ...

{pstd}
or you type

	{cmd:. sts graph} ...

{pstd}
the result is that those commands construct a complicated
{cmd:graph} {cmd:twoway} command

	{cmd:-> graph twoway} {it:something_complicated}

{pstd}
and then run that for you.  When you specify the addplot option, such as

	{cmd:. histogram} ...{cmd:, addplot(}{it:your_contribution}{cmd:)}

{pstd}
or

	{cmd:. sts graph, addplot(}{it:your_contribution}{cmd:)}

{pstd}
the result is that the commands construct

{phang2}
	{cmd:-> graph twoway} {it:something_complicated} {cmd:||}
		{it:your_contribution}

{pstd}
Let us assume that you have survival data and wish to visually compare the
Kaplan-Meier (i.e., the empirical survivor function) to the function that
would be predicted if the survival times were assumed to be exponentially
distributed.  Simply typing

	{cmd:. sysuse cancer, clear}

	{cmd:. quietly stset studytime, fail(died)}

	{cmd:. sts graph}
	  {it:({stata "gr_example2 plotop1":click to run})}
{* graph plotop1}{...}

{pstd}
will obtain a graph of the empirical estimate.
To obtain the exponential estimate, you might type

	{cmd:. quietly streg, distribution(exponential)}

	{cmd:. predict S, surv}

	{cmd:. graph twoway line S _t, sort}
	  {it:({stata "gr_example2 plotop2":click to run})}
{* graph plotop2}{...}

{pstd}
To put these two graphs together, you can type

	{cmd:. sts graph, addplot(line S _t, sort)}
	  {it:({stata "gr_example2 plotop3":click to run})}
{* graph plotop3}{...}

{pstd}
The result is just as if you typed

	{cmd:. sts graph || line S _t, sort}

{pstd}
if only that were allowed.


{title:Also see}

{psee}
Manual:  {bf:[G] {it:addplot_option}}

{psee}
Online:  {helpb twoway}
{p_end}