compare prefab

Version 2.42 May2013


View the "prefab compare" examples

As of version 2.40 this prefab is currently suspended and nonfunctional. support for it is suspended... I'm not sure if there's any user interest out there... if there is, this prefab needs to be rewritten, probably to use proc processdata action: summary (the old proc rangebar functionality is no longer supported). Community involvement in accomplishing this is solicited.

compare is a prefab designed to allow convenient comparison of one to four variables. It computes means and standard deviations, and then produces groups of bars with error bars, arranged for easy comparison. Categories and subcategories are supported, and from one to four data fields/variables can be examined. I wrote it because I wanted an easy way to compare 2, 3, or 4 measurments for different mouse strains (categories), with female and male separate (subcategories). This prefab does not support more than 2 subcategories. This prefab is new in version 2.21-01.


Data type DefaultDescription Examples
(ignore line wrapping)
Standard parameters     Any of the standard prefab parameters may be used. You must specify data (the name of the input data file).

data field y must be specified Specifies the data field(s) to be displayed and compared. y=Rbw y2=Lbw
catfields data field, or comma-delimited list of 2 data fields. must be specified One field that identifies the category, or two data fields identifying the category and the subcategory, for each data row. The data set must be sorted on these fields. catfields=2
leftsub text must be specified if doing subgroups A value that will be found in the subgroup field (the 2nd catfield value as described above) of data records. Data records having this tag will be considered part of the subgroup that is displayed on the left. For example, if the subgroups are denoted in the data using tags of female and male, and female subgroup appears on the left, then use leftsub=female. See examples 4 and 5. leftsub=female
rightsub text must be specified if doing subgroups Same as leftsub but for the subgroup appearing on the right. rightsub=male
barwidth inches or cm 0.05 " Width of bars. barwidth=0.08
subgroupoffset Units in X 0.4 Distance between clusters when subgroups are being displayed. subgroupoffset=0.5
clustersep inches or cm 0.012 " Amount of separation between individual bars within clusters clustersep=0

color orange, powderblue, green, dullyellow Colors for the bars. color goes with y, color2 goes with y2, and so on. color=red color2=green
outline yes
  If yes, outline the bars. outline=yes
crossover value in Y   This may be specified as the value in Y where bars begin to point downward instead of upward. For instance, crossover=0 is often useful when dealing with financial income, to show loss using downward bars. crossover=0

select expression   Specifies conditionals for selecting data rows. For instance, can be used so that one data field is used for y, y2 (etc.), with the segregation being performed by the conditional expressions. See example 6. ptselecttags must also be specified for this to work. Note that since these are passed in only one at-sign should be used, not two. "ptselect = 24 && = ctrl"
ptselecttags comma-delimited list must be specified if ptselect* used This is a comma-delimited list of tags or identifiers, one unique member per ptselect. For example, if you are using ptselect and ptselect2 then the ptselecttags list must have two members. The tags will be shown in the default legend, and also be used internally. See example 6. ptselecttags="0,3"
dataoutfile filename   If specified, computed summary statistics will be written to this file. Fields are: the catfields (1 or 2), variable name, N, mean, SD, median, min, max, and number of missing (non-numeric) observations. dataoutfile=tmp1
errline linedetails width=0.4 Controls the thickness, color, etc. of all error bars. "errline= width=1"
errwidth inches or cm 0.05" Size of error bar tails errwidth=0.8
erroneway yes or no no If yes, error bars are only drawn one way (upward). erroneway=yes

data display engine  
Copyright Steve Grubb

Ploticus is hosted at   Get ploticus data display engine at Fast, secure and Free Open Source software downloads