Evaluating sample pairs using InfoSphere MDM Pair Manager
To establish thresholds that accurately match members, you must evaluate the sample pairs file that was created as part of the weight generation process. InfoSphere® MDM Pair Manager presents each pair side-by-side, and highlights each matching row of data. By using InfoSphere MDM Pair Manager, you can quickly determine whether the two members of each pair match, with minimal scrolling.
Before you begin
- Ensure that there is a suitable IBM Java 1.7 or 1.8 JRE installed on the system.Note: A suitable JRE is installed with both WebSphere Application Server and Rational Application Developer.
- Ensure that the system PATH environment variable is updated to include the bin folder of the suitable JRE, such as <RAD_INSTALL_HOME>\jdk\jre\bin.
- Open a command prompt window and change the current directory to the InfoSphere MDM Pair Manager installation folder under the MDM_INSTALL_HOME folder, such as MDM_INSTALL_HOME\pairmanager\lib.
Launch InfoSphere MDM
Pair Manager using the
java -jar com.ibm.mdm.pairmanager.jar
- Open the samplePairs.xls file by clicking the Open button.
- In the Open Sample Pair File window,
browse to the files location and click Open.
The samplePairs.xls file is typically in Workbench
project_namedirectory, but might be saved to another path. (The file name you open might differ from the default samplePairs.xls). The InfoSphere MDM Pair Manager screen is populated with pair data for the attributes selected in the Generate Threshold Analysis Pairs job.
Matching attributes are marked with a green check mark in the Pair Data Labels column. The matching attribute data is highlighted in green for both members. Some members might have multiple attributes, such as the last name.
- For each pair of members, evaluate each data
element to determine whether the members are the same.
- If they are the same, click the checkmark (is a match) button or press the Y key.
- If they are not the same, click the slashed circle (is not a match) button or press the N key.
- If it cannot be determined from the data shown, click the question mark (might be a match) button or press the M key.
- To advance to the next pair, click the Right arrow (next) button or press the Right Arrow key. If the Auto-advance option is enabled, the next pair is displayed automatically. If you need to return to the previous pair, click the Left arrow (previous) button or press the Left Arrow key.
- When finished evaluating pairs, click the Save button. You can overwrite the file or specify a new file name.
If you do not evaluate all pairs in one sitting, you can use the InfoSphere MDM Pair Manager again later to continue where you left off. Use the filtering options to easily locate the pairs that still require evaluation.