Search
Ploticus >
Scripts >
| xautorange and yautorange attributes |
Usage
xautorange: datafield=dfields other settings..
or..
yautorange: datafield=dfields other settings..
Data-driven axis range determination
xautorange and yautorange are attributes of
proc areadef
that allow axis range to be automatically determined based on the range of the data.
See also gallery examples
scatterplot10,
stockday,
and
mouse.
Examples
Example: xautorange: datafield=2 nearest=month margin=31
Example: xautorange: datafield=1 nearest=monday
Example: yautorange: datafield=3,4,5,6,7 selectrows=@@sex = f
Settings
The datafield setting must be specified, and it must be leftmost.
All other settings are optional and should default to reasonable behavior.
datafield=dfields
This setting must be given first (to the left of any other).
It specifies one or more data fields that will be scanned to
automatically find minima and maxima.
dfields may be a single
dfield
or a comma-delimited list of several dfields with no embedded space.
Unless a combomode= setting is used, all fields will be treated independently
when searching for a minima and maxima.
Invalid or non-existant dfields will be silently ignored, but there must be at least one valid dfield.
Support for multiple fields contributed by Paul Totten.
Example: yautorange datafield=2
Example: yautorange datafield=2,4,5
combomode=stack | combomode=hilo
If specified, an alternate method of min/max determination is used with multiple datafields.
stack = sum the data field contents (useful when generating stacked bar graphs).
hilo = take the sum and difference of data field contents; useful when
ranging on two fields, and one data field will be driving error bars.
nearest=s
Set the min and max to fall on round boundaries.
For instance, nearest=10 can be used to ensure that the axis min and max will always fall at
an even multiple of 10. nearest=month ensures that the axis min and max will fall on an even
month boundary.
With numeric data, s can be any number greater than zero.
For date, time, and datetime data s is a controlled vocabulary (see below).
If plotting numeric data and nearest is not specified, a reasonable s will be selected.
If plotting date, time, or datetime data and nearest is not specified, datematic will be used
(previously to 2.31 exact was the default).
Example: nearest=100 and the data range from 23 to 1162, the plot's range will be 0 to 1200.
Possible values for nearest are shown below.
For each, the applicable data types are shown in parentheses.
nearest= range min and max will be set to:
-------- ---------------------------------
0.1 the nearest multiple of 0.1 (numeric data)
10 the nearest multiple of 10 (numeric)
1000 the nearest multiple of 100 (numeric)
(etc... any number value can be used for numeric data)
exact the exact min and max as found in the data (any data type)
datematic a reasonable default will be used (date, time, datetime)
minute the nearest 1 minute boundary (time data)
Nminute the nearest N minute boundary (time data)*
second the nearest 1 second boundary (time data)
Nsecond the nearest N second boundary (time data)*
hour the nearest 1 hour boundary (time, datetime)
3hour the nearest 3 hour boundary (time, datetime)
6hour the nearest 6 hour boundary (time, datetime)
12hour the nearest 12 hour boundary (time, datetime)
day the nearest day boundary (datetime)
sunday the nearest Sunday (date, datetime)
monday the nearest Monday (date, datetime)
month the nearest month boundary (date, datetime)
quarter the nearest 3 month boundary (date, datetime)
3month same as quarter
year the nearest year boundary (date, datetime)
2year the nearest 2 year boundary (date, datetime)
5year the nearest 5 year boundary (date, datetime)
10year the nearest 10 year boundary (date, datetime)
* N is a one or two digit integer and must go into 60 evenly
margin=m
If specified, m basic units will be added to the
maxima and also subtracted from the minima, before considering nearest.
This may be used to guarantee margins at each end of the range.
(Basic units are days for date and datetime scaletypes; and
minutes for time.)
incmult=h
a more automatic way to get the equivalent of margin with numeric data.
If specified, the default scale increment (same as used by proc axis stubs: inc)
will be used for the margin. h can be set to 1 to use the increment
as is, or it can be set to a value such as 2 to double it.
The default h is 1.0.
The low end of the range will not dip below zero if data are all positive.
Example: yautorange datafield=2 incmult=2.0
lowfix=lowval
If specified, the minima will always be lowval.
This is useful when you want only the maxima to be automatically set based on data.
Example: yautorange: datafield=2 lowfix=0
hifix=hival
If specified, the maxima will always be hival.
This is useful when you want only the minima to be automatically set based on data.
mininit=lowval
Like lowfix but is overridden by encountered data values less than lowval.
Useful for example if you want your axis min to be zero except when there is an occasional
negative data value. (2.30+)
maxinit=hival
Like hifix but is overridden by encountered data values greater than hival. (2.30+)
numformat=printfspec
This allows user control over the internal string-based representation of minima and maxima for numerics.
Default is %g which handles scientific-notation values and non-extreme numerics.
Try %f if your data include extreme numerics that aren't in scientific notation.
selectrows=conditional expression
This allows a
select expression
to be applied in order to select data rows to be tested when
scanning for minima and maxima.
If specified, this setting must be the last (rightmost) on the line, and
will have embedded spaces, unlike the other settings. See example below.
Variables that are set by autorange
DATAXMIN, DATAXMAX
The actual minima and maxima of the data field when ranging in X.
Different from XMIN and XMAX in that these do not include any margin.
DATAYMIN, DATAYMAX
The actual minima and maxima of the data field when ranging in Y.
Different from YMIN and YMAX in that these do not include any margin.
|
