Determines the data limits for an array using a variety of methods KPG1_DARAx
The method is selected by the user through the literal parameter specified by PARAM. The supplied string should consist of up to three sub-strings or elements, separated by commas. In all but one method, the first element must specify the method to use. If supplied, the other two elements should be numerical values as described below (default values will be used if these elements are not provided). The following options are available.
" Extended" – LOWER and UPPER are returned equal to the lowest and highest supplied data values,
extended to give a margin of 2.5% of the total data range at each end.
" Extended,10,5" – Like " Extended" , except the margins at the two ends are specified as a pair of
numerical value in the second and third elements of the array. These values are percentages of the
total data range. So, " Extended,10,5" includes a margin of 10% of the total data range in LOWER, and
5% in UPPER. If only one numerical value is given, the same value is used for both limits.
If no value is given, both limits default to 2.5. " Range" is equivalent to " Extended,0,0"
.
" Faint" – The scaling is between the mean data value minus one standard deviation and the mean
data value plus seven standard deviations. This is retained for historical reasons.
" Limit" – Allows user to supply explicit lower and upper limiting values. For example, " Limit,10,20"
would set the lower limit to 10 and its upper limit to 20. For compatibility reasons and for ease of use,
this routine permits the omission of the method. Thus " 10,20" is equivalent to the previous example.
The upper limit may be numerically smaller than the lower limit, as might be needed for
display scaling. If fewer than two numerical values are supplied, the " Range" method is
adopted. The limits must be within the allowed range for the data type of the ARRAY
argument.
" Percentiles,5,95" – The second and third elements of the array are interpreted as percentiles. For
instance, " Perc,5,95" causes 5% of the data points to be below LOWER, and 10% to be above the
UPPER. If only 1 value (p1) is supplied, the other one, p2, defaults to (100 - p1). If no values
are supplied, p1 and p2 default to 5 and 95 respectively. Values must be in the range 0 to
100.
" Range" – LOWER and UPPER are returned equal to the lowest and highest supplied array
values.
" Sigmas,2,3" – The second and third elements of the array are interpreted as multiples of the standard
deviation of the data values. For instance, " S,2,3" causes the LOWER to be the mean of the data
values, minus two sigma, and UPPER to be the mean plus three sigma. If only 1 value is supplied,
the same value is used for both limits. If no values are supplied, both values default to
3.0.
The above method names can be abbreviated to one character.
If the parameter name is supplied as blank, then " Range" is assumed (i.e. LOWER and UPPER are
chosen so that the encompasses the entire data range).
The extreme values are reported unless the limits parameter PARAM is specified on the command line. In this case extreme values are only calculated where necessary for the chosen method.
" Faint" , "
Extended" , " Limit" , " Percentiles" , " Range" , and " Sigmas" . See the description for details of
these methods. There is a routine for the all numeric data types: replace " x" in the routine name by B, D, I, R, UB,
UW, or W as appropriate. The ARRAY argument supplied must have the data type specified.