This method sets outlier limits for the specified field contained
in a DM_LogicalDataSpec value. All values that fall outside
of this range are treated as outliers. This applies only to
numeric fields.
Parameters
- logData
- A value of type DM_LogicalDataSpec
- fldName
- A value of type VARCHAR(128) that is already contained
in logData
- lowLimit
- A value of type DOUBLE , DATE, TIME, or TIMESTAMP, that
represents the lower bound value
- highLimit
- A value of type DOUBLE , DATE, TIME, or TIMESTAMP, that
represents the upper bound value
Return Value
The return value is of type DM_LogicalDataSpec
- If fldName is NULL, the value logData is returned.
- If logData does not contain fldName, the following
error occurs:
IDM1201E: A field with the name "Variable1" is not
available.
- If fldName is of type DM_Categorical
, the following error occurs:
IDM1206E: The field "Variable1"
is not numeric.
- If lowLimit is greater than highLimit or lowLimit equals highLimit,
the following error occurs:
IDM1207E: The lower bound must be smaller
than the upper bound.
- If lowLimit is NULL or highLimit is NULL, the value logData is
returned. Lower and upper limits for field values are
not set. During the training run, lower and upper limits
for field values will be calculated by the algorithm according
to the values in the data.
- If none of the conditions above apply the value logData is
returned. The field with a name equal to fldName is
updated with the given values. If the values are of type
DATE, TIME, or TIMESTAMP, they are first converted to
a DOUBLE value.