Here is one approach:
Names Default to Here( 1 );
dt = Open( "$SAMPLE_DATA/Big Class.jmp" );
Summarize( g = By( :sex ) );
If( N Items( g ) < 2,
biv obj = dt << Bivariate( Y( :weight ), X( :height ), Fit Line );
biv rep = biv obj << Report;
reg est = biv rep["Parameter Estimates"][NumberColBox(1)] << Get As Matrix,
// need By variable
biv obj = dt << Bivariate( Y( :weight ), X( :height ), By( :sex ), Fit Line );
biv rep = biv obj << Report;
reg est = [];
For( by = 1, by <= N Items( g ), by++,
reg est ||= (biv rep[by])["Parameter Estimates"][NumberColBox(1)] << Get As Matrix;
);
);
Show( reg est );
Hello....wondering if someone can explain the following discrepancy with the number of items in a report. I have the following code which is trying to extract the intercept and slope out of a linear fit that has been applied to a Fit Y by X platform. The issue I see is if the :group column contains just one item then the numplots = 2, and I get an error trying to extract the intercept and slope from rp[1]. I have to extract from rp[2]. If the number of items in the :group column = 2 or higher then numplots = the number of items in the group column and I can extract intercept and slope from rp[1]. In the end, I want to be able to extract all intercepts and slopes for x number of groups where x can be anything from 1 to whatever, but this "feature" makes coding a bit more long winded because I have to treat # groups = 1 different than # goups >= 2.
Clear Globals();
dt = Current Data Table();
bivObj = dt << Bivariate(
Y(:this),
X(:that),
Fit Line,
By(:group),
);
rp = bivObj << Report;
numPlots = nItems(rp1);
intercept = rp[1][OutlineBox("Parameter Estimates")][Columnbox("Estimate")][1];
slope = rp[1][OutlineBox("Parameter Estimates")][Columnbox("Estimate")][1];
thanks
When you use a By() in a platform call if there is more than one group the result will be a list of the platforms.
Consider the following:
dt = Open( "$SAMPLE_DATA\Big Class.jmp" );
dt <<
New Column( "Constant",
character,
formula( "a" )
);
bivObj1 = dt <<
Bivariate(
Y( :height ),
X( :weight ),
Fit Line,
By( :sex ),
);
bivObj2 = dt <<
Bivariate(
Y( :height ),
X( :weight ),
Fit Line,
By( :constant ),
);
Show( bivobj1, bivobj2 );
The log will show:
/*:
bivobj1 = {Bivariate[], Bivariate[]};
bivobj2 = Bivariate[];
So, you can check your bivObj variable to see if it's a list and act accordingly. For example:
dt = Open( "$SAMPLE_DATA\Big Class.jmp" );
dt << New Column( "Constant", character, formula( "a" ) );
bivObj1 = dt << Bivariate(
Y( :height ),
X( :weight ),
Fit Line,
By( :sex ),
);
bivObj2 = dt << Bivariate(
Y( :height ),
X( :weight ),
Fit Line,
By( :constant ),
);
Show( bivobj1, bivobj2 );
If( Is List( bivobj1 ),
//bivobj1 is a list of platform references
For( i = 1, i <= N Items( bivobj1 ), i++,
rp = bivObj1[i] << Report;
intercept = rp[Outline Box( "Parameter Estimates" )][Columnbox( "Estimate" )][1];
slope = rp[Outline Box( "Parameter Estimates" )][Columnbox( "Estimate" )][1];
Show( intercept, slope );
)
,
//bivobj1 is not a list, it's a single platform reference
rp = bivObj1 << Report;
intercept = rp[Outline Box( "Parameter Estimates" )][Columnbox( "Estimate" )][1];
slope = rp[Outline Box( "Parameter Estimates" )][Columnbox( "Estimate" )][1];
);