relate

is a batch search application that accesses the IIR Search Server using the standard Search APIs.

It reads an input file containing search transactions. Each search transaction is passed to the Search Server which uses the nominated Search Definition to find matching records. These are written to the output file by relate.

Input records must be separated by a newline. By default their format must match the layout of the IDT to be searched. If the format differs from the IDT layout, the

-i

switch can be used to nominate an input format. Multiple output formats are supported. These are controlled with the

-o

switch.

Normally all records returned by the Search Server are written to the output file. That is those records that have an acceptable score as determined by the Search Definition. Additional filtering is possible with the

-l

and

-u

switches. These are used to set upper and lower bounds for acceptable scores and are applied by relate prior to emitting records to the output report. Note that filtering with

-l

and

-u

is not integrated with the other options. The filter will remove all records that are not within bounds, even if this will result in an empty set, irrespective of whether or not non-empty sets were requested. For example with the

-x

or

-m

switches. These filters are primarily used to experiment with score thresholds. Once correct thresholds have been determined, add them to the Search Definition and discontinue the use of

-l

and

-u

.

The

-m

switch can be used to create multiple output files.

relate

can also search for duplicate records in the IDT. When started with the

-x

switch, relate runs in DupFinder mode.