You are here: Processor Library > Matching > Output Selectors > Earliest Value

Output selector: Earliest Value

The Earliest Value output selector evaluates a date stamp on each record to select a value to use for another attribute.

Use

Use the Earliest Value output selector where records have date stamps, and the best value for an attribute is likely to come from the record with the earliest stamp.

For example, in some cases, data is carefully checked and cleaned on migration to a new system, but duplicate entries are created with errors. In this case, the earlier records may be better. The date stamp column must be selected as an input to the output selector, in addition to the attribute for which you wish to determine a merged value.

Inputs

For the actual output value, any input attributes of any type from any input data sets may be configured.

A Date attribute must be configured in order to establish which record should be considered the 'Earliest'.

Options

Option

Type

Purpose

Default Value

Use first non-empty value if tied?

Yes/No

This option provides a way of selecting a value arbitrarily if more than one record exists with the same, earliest, Date/time stamp.

The first alphabetically sorted value will be selected for String values, the lowest value for Numbers, and the earliest value for Dates.

Yes

Example

In this example, the Earliest Value output selector is used to select the Earliest Value for a Company Name field, based on the date stamps of the records.

Example configuration

Use first non-empty value if tied? = No

Example output

The following table illustrates some example output using the Earliest Value selector:

Input A (CompanyName, Date)

Input B (CompanyName, Date)

Output value (Earliest Value)

Barclays Bank plc,

10/01/1998

Barclays Bank (Bristol) PLC,

14/05/2002

Barclays Bank plc

PriceWaterhouse Coopers,

10/01/1998

PWC,

24/03/2000

PriceWaterhouse Coopers

Oracle Limited,

24/03/2003

Oracle,

24/03/2003

Selection error (needs manual resolution)

Oracle, null

Oracle, 24/01/1997

Oracle

Oracle ® Enterprise Data Quality Help version 9.0
Copyright © 2006,2011 Oracle and/or its affiliates. All rights reserved.