graph_dot.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

    has an effect only when {cmd:if} or {cmd:in} are specified or if there are
    missing values in the variables.  {cmd:allcategories} may not be combined
    with {cmd:by()}.


{title:yvar_options}

{phang}
{cmd:ascategory}
    specifies that the {it:yvars} be treated as the first {cmd:over()} group.

{pmore}
    When you specify {cmd:ascategory}, results are the same as if you
    specified a single {it:yvar} and introduced a new first {cmd:over()}
    variable.  Anyplace you read in the documentation that something is
    done over the first {cmd:over()} category, or using the first
    {cmd:over()} category, it will be done over or using {it:yvars}.

{pmore}
    Suppose that you specified

{phang3}
	    {cmd:. graph dot y1 y2 y3, ascategory} {it:whatever_other_options}

{pmore}
    The results will be the same as if you typed

{phang3}
	    {cmd:. graph dot} {it:y}{cmd:, over(}{it:newcategoryvariable}{cmd:)} {it:whatever_other_options}

{pmore}
    with a long rather than wide dataset in memory.

{phang}
{cmd:asyvars}
    specifies that the first {cmd:over()} group be treated as {it:yvars}.

{pmore}
    When you specify {cmd:asyvars}, results are the same as if you removed
    the first {cmd:over()} group and introduced multiple {it:yvars}.  Note
    that we said in most ways, not all ways, but let's ignore that for a
    moment.  If you previously had {it:k} {it:yvars} and, in your first
    {cmd:over()} category, {it:G} groups, results will be the same as if you
    specified {it:k}*{it:G} yvars and removed the {cmd:over()}.  Anyplace you
    read in the documentation that something is done over the {it:yvars} or
    using the {it:yvars}, it will be done over or using the first {cmd:over()}
    group.

{pmore}
    Suppose that you specified

{phang3}
	    {cmd:. graph dot y, over(group) asyvars} {it:whatever_other_options}

{pmore}
    Results will be the same as if you typed

{phang3}
	    {cmd:. graph dot} {it:y1 y2 y3} ...{cmd:,} {it:whatever_other_options}

{pmore}
    with a wide rather than long dataset in memory.
    Variables {it:y1}, {it:y2}, ..., are sometimes called the virtual
    {it:yvars}.

{phang}
{cmd:percentages}
    specifies that marker positions be based on percentages that
    {it:yvar}_{it:i} represents of all the {it:yvars}.  That is,

{phang3}
	    {cmd:. graph dot (mean) inc_male inc_female}

{pmore}
    would produce a chart with the markers reflecting average income.

{phang3}
	    {cmd:. graph dot (mean) inc_male inc_female, percentage}

{pmore}
    would produce a chart with the markers being located at
    100*inc_male/(inc_male+inc_female) and
    100*inc_female/(inc_male+inc_female).

{pmore}
    If you have a single {it:yvar} and want percentages calculated over the
    first {cmd:over()} group, specify the {cmd:asyvars} option.  For instance,

{phang3}
	    {cmd:. graph dot (mean) wage, over(}{it:i}{cmd:) over(}{it:j}{cmd:)}

{pmore}
    would produce a chart where marker positions reflect mean wages.

{phang3}
	    {cmd:. graph dot (mean) wage, over(}{it:i}{cmd:) over(}{it:j}{cmd:) asyvars percentages}

{pmore}
   would produce a chart where marker positions are

	    100*( mean_ij / (Sum_i mean_ij) )

{phang}
{cmd:cw}
    specifies casewise deletion.  If {cmd:cw} is specified, observations for
    which any of the {it:yvars} are missing are ignored.  The default is
    to calculate each statistic using all the data possible.


{title:linelook_options}

{phang}
{cmd:outergap(*}{it:#}{cmd:)} and
{cmd:outergap(}{it:#}{cmd:)}
    specify the gap between the top of the graph to the beginning of the first
    line and the last line to the bottom of the graph.

{pmore}
    {cmd:outergap(*}{it:#}{cmd:)} specifies that the default be modified.
    Specifying {cmd:outergap(*1.2)} increases the gap by 20%, and
    specifying {cmd:outergap(*.8)} reduces the gap by 20%.

{pmore}
    {cmd:outergap(}{it:#}{cmd:)} specifies the gap as a
    percentage-of-bar-width units.  {cmd:graph} {cmd:dot} is related to
    {cmd:graph} {cmd:bar}.  Just remember that {cmd:outergap(50)} specifies a
    sizable but not excessive gap.

{phang}
{cmd:linegap(}{it:#}{cmd:)}
    specifies the gap to be left between {it:yvar} lines.
    The default is {cmd:linegap(0)}, meaning that multiple {it:yvars} appear
    on the same line.  For instance, typing

	    {cmd:. graph dot y1 y2, over(group)}

{pmore}
    results in

		group 1  {c |}..x....o........
		group 2  {c |}........x..o....
		group 3  {c |}.......x.....o..
			 {c BLC}{c TT}{hline 4}{c TT}{hline 4}{c TT}{hline 4}{c TT}

{pmore}
    In the above, o represents the symbol for y1 and x the symbol for y2.
    If you want to have separate lines for the separate
    {it:yvars}, specify {cmd:linegap(20)}:

	    {cmd:. graph dot y1 y2, over(group) linegap(20)}

		group 1  {c |}.......o........
			 {c |}..x.............
			 {c |}
		group 2  {c |}...........o....
			 {c |}........x.......
			 {c |}
		group 3  {c |}.............o..
			 {c |}.......x........
			 {c BLC}{c TT}{hline 4}{c TT}{hline 4}{c TT}{hline 4}{c TT}

{pmore}
    Specify a number smaller or larger than 20 to reduce or increase the
    distance between the y1 and y2 lines.

{pmore}
    Alternatively, and generally preferred, is specifying option
    {cmd:ascategory}, which will result in

	    {cmd:. graph dot y1 y2, over(group) ascategory}

		group 1  y1  {c |}.......o........
			 y2  {c |}..o.............
			     {c |}
		group 2  y1  {c |}...........o....
			 y2  {c |}........o.......
			     {c |}
		group 3  y1  {c |}.............o..
			 y2  {c |}.......o........
			     {c BLC}{c TT}{hline 4}{c TT}{hline 4}{c TT}{hline 4}{c TT}


{pmore}
    Note that {cmd:linegap()} affects only the {it:yvar} lines.  If you want
    to change the gap for the first, second, or third {cmd:over()}
    groups, specify the {it:over_subopt} {cmd:gap()} inside the
    {cmd:over()} itself.

{phang}
{cmd:marker(}{it:#}{cmd:,} {it:marker_options}{cmd:)}
    specifies the shape, size, color, etc., of the marker to be used to mark
    the value of the {it:#}th {it:yvar} variable.  {cmd:marker(1,} ...{cmd:)}
    refers to the marker associated with the first {it:yvar}, {cmd:marker(2,}
    ...{cmd:)} refers to the marker associated with the second, and so on.  A
    particularly useful {it:marker_option} is
    {cmd:mcolor(}{it:colorstyle}{cmd:)}, which sets the color of the marker.
    For instance, you might specify {cmd:marker(1, mcolor(green))} to make the
    marker associated with the first {it:yvar} green.  See
    {it:{help colorstyle}} for a list of color choices, and see 
    {it:{help marker_options}} for information on the other
    {it:marker_options}.

{phang}
{cmd:pcycle(}{it:#}{cmd:)}
    specifies how many variables are to be plotted before the {help pstyle} of
    the markers for the next variable begins again at the {cmd:pstyle} of the
    first variable {c -} {cmd:p1dot} (with the markers for the variable
    following that using {cmd:p2dot} and so).  Put another way, {it:#}
    specifies how quickly the look of markers is recycled when more than
    {it:#} variables are specified.  The default for most {help schemes}
    is {cmd:pcycle(15)}.

{phang}
{cmd:linetype(dot)},
{cmd:linetype(line)}, and
{cmd:linetype(rectangle)}
    specify the style of the line.

{pmore}
    {cmd:linetype(dot)} is the usual default.  In this style, dots are used
    to fill the line around the marker:
		       ........o........

{pmore}
    {cmd:linetype(line)} specifies that a solid line be used to fill the
    line around the marker:
		       {hline 8}o{hline 8}

{pmore}
    {cmd:linetype(rectangle)} specifies that a long "rectangle" (which looks
    more like two parallel lines) be used to fill the area around the marker:
		       ========o=======

{phang}
{cmd:ndots(}{it:#}{cmd:)} and
{cmd:dots(}{it:marker_options}{cmd:)}
    are relevant only in the {cmd:linetype(dots)} case.

{pmore}
    {cmd:ndots(}{it:#}{cmd:)} specifies the number of dots to be used to fill
    the line.  The default is {cmd:ndots(100)}.

{pmore}
    {cmd:dots(}{it:marker_options}{cmd:)} specifies the marker symbol, color,
    and size to be used as the dot symbol.  The default is to use
    {cmd:dots(msymbol(p))}.  See {it:{help marker_options}}.

{phang}
{cmd:lines(}{it:line_options}{cmd:)}
    is relevant only if {cmd:linetype(line)} is specified.  It specifies the
    look of the line to be used; see {it:{help line_options}}.

{phang}
{cmd:rectangles(}{it:area_options}{cmd:)} and
{cmd:rwidth(}{it:relativesize}{cmd:)}
    are relevant only if {cmd:linetype(rectangle)} is specified.

{pmore}
    {cmd:rectangles(}{it:area_options}{cmd:)}
    specifies the look of the parallel lines (rectangle);
    see {it:{help area_options}}.

{pmore}
    {cmd:rwidth(}{it:relativesize}{cmd:)}
    specifies the width (height) of the rectangle (the distance between
    the parallel lines).  The default is usually {cmd:rwidth(.45)};
    see {it:{help relativesize}}.

{phang}
{cmd:noextendline} and {cmd:extendline}
    are relevant in all cases.  They specify whether the line (dots, a
    line, or a rectangle) is to extend through the plot region margin and 
    touch the axes.  The usual default is {cmd:extendline}, so
    {cmd:noextendline} is the option.  See {it:{help region_options}} for
    a definition of the plot region.

{phang}
{cmd:lowextension(}{it:relativesize}{cmd:)}
and
{cmd:highextension(}{it:relativesize}{cmd:)}
    are advanced options that specify the amount by which the line (dots, line
    or a rectangle) is extended through the axes.  The usual defaults are
    {cmd:lowextension(0)} and {cmd:highextension(0)}.  See
    {it:{help relativesize}}.


{title:legending_options}

{phang}
{it:legend_option}
    allows you to control the legend.  If more than one {it:yvar} is
    specified, a legend is produced.  Otherwise, no legend is needed because
    the {cmd:over()} groups are labeled on the categorical {it:x} axis.  See
    {it:{help legend_option}}.

{phang}
{cmd:nolabel}
    specifies that, in automatically constructing the legend, the variable
    names of the {it:yvars} be used in preference to "mean of {it:varname}" or
    "sum of {it:varname}", etc.

{phang}
{cmd:yvaroptions(}{it:over_subopts}{cmd:)}
    allows you to specify {it:over_subopts} for the {it:yvars}.  This is
    very rarely done.

{phang}
{cmd:showyvars}
    specifies that, in addition to building a legend, the identities of the
    {it:yvars} be shown on the categorical {it:x} axis.  If {cmd:showyvars} is
    specified, it is typical to also specify {cmd:legend(off)}.


{title:axis_options}

{phang}
{cmd:yalternate} and {cmd:xalternate}
    switch the side on which the axes appear.
    {cmd:yalternate} moves the numerical
    {it:y} axis from the bottom to the top; {cmd:xalternate} moves the
    categorical {it:x} axis from the left to the right.
    If your scheme by default puts the axes on the opposite sides,
    {cmd:yalternate} and {cmd:xalternate} reverse their actions.

{phang}
{cmd:exclude0}
    specifies that the numerical {it:y} axis need not be scaled to include 0.

{phang}
{cmd:yreverse}
    specifies that the numerical {it:y} axis have its scale reversed
    so that it runs from maximum to minimum.

{phang}