Oracle Cloud Learning Center


19 Manage Product and Service Data Quality: Cleanse Product and Service Data

This button toggles the Table of Contents floating window

 

This chapter contains the following:

Check Data Quality

FAQs for Check Data Quality

Check Data Quality

Data Quality: Explained

The quality of product data is enhanced by integration with Oracle Product Data Quality, which provides classification, standardization, and matching to refine item data and to prevent duplicate items. Item data can be inconsistent, especially when imported from external sources.

The integration with Oracle Product Data Quality involves the following:

  • Data quality checking

  • The semantic model

  • Data quality attributes

  • Classification

  • Standardization

  • Matching

Data Quality Checking

When you check data quality, Oracle Product Data Quality applies classification, standardization, and matching to the items that you are creating, editing, or importing, and then presents you with the results for acceptance.

You can check the quality of your product data:

  • When you are creating a single item interactively

  • When you are creating multiple items interactively

  • When you are editing an item interactively

  • By running a periodically scheduled process that checks the items in an item class

  • When you are importing a batch of items

The Semantic Model

Before you can check data quality in Oracle Fusion Product Hub, you must initially set up the semantic model in Oracle Product Data Quality. The semantic model is a fully configured data lens, which contains detailed contextual information about your data. The semantic model contains your definitions for classification, standardization, and matching.

You can help build up the initial semantic model by running a scheduled process that extracts metadata from the item class and catalog tables of Oracle Fusion Product Hub. The use of metadata in setting up the semantic model is described in the documentation for Oracle Product Data Quality about AutoBuild.

You can also provide the AutoBuild feature with sample item data by exporting the results of an item search into an external spreadsheet, then transferring that spreadsheet data into Oracle Product Data Quality. To export the data, perform a search on the Manage Items page, then in the Search Results section, select Actions - Export to Excel.

Data Quality Attributes

Data quality depends primarily on the values of designated attributes. In Oracle Product Data Quality, you define the detailed rules for the relationships between these data quality attributes, which are used for either standardization or matching. In Oracle Fusion Product Hub, you designate these attributes at the item class level.

Classification

In Oracle Product Data Quality, you define rules that classify data items as belonging to the appropriate item class based on the values of certain attributes. Classification also includes suggested assignments of items to one or more catalogs and to categories within those catalogs. For example, if the Power attribute equals 10 watts and the Capacity attribute equals 300 ohms, then the item belongs to the Capacitor item class.

Standardization

In Oracle Product Data Quality, you define rules that make the values of specified item attributes consistent with desired norms. For example, you might convert all Fahrenheit temperature values to Celsius, or all English measurements to metric. You can also use standardization to merge divergent forms of attribute values into a single consistent form, such as by changing the unit of measure values in. and IN to Inches.

Matching

In Oracle Product Data Quality, you define rules for detecting when an item that is being created or imported matches an item that already exists in the Oracle Fusion Product Hub repository. For example, suppose that if power supply items differ only in the finish style of the casing, then they are considered to match existing items. To get this result, you would not use the hypothetical Casing Style attribute as a match-rule attribute.

Check Data Quality for Items: Examples

The following scenarios illustrate the ways in which you can check data quality when working with items.

  • Creating a single item

  • Creating multiple items

  • Editing an item

  • Running a scheduled process

  • Importing an item batch

Creating a Single Item

To check data quality when creating a single item interactively:

  1. On the Create Item page, select Actions - Check Data Quality.

  2. Examine the Check and Apply Data Quality Results window, which displays the classification, standardization, and matching values provided for the item by Oracle Product Data Quality.

  3. If the data quality results are satisfactory, click Apply to apply the new values to the Create Item page.

  4. The data quality checks are also performed when you click Save, Save and Close, or Submit. Note that you can submit an item only when its item class is enabled for new item requests.

Creating Multiple Items

To check data quality when you are creating multiple items interactively:

  1. On the Create Multiple Items page, select Actions - Check Data Quality.

  2. Examine the Check and Apply Data Quality Results window, which displays, for each item in the set, the classification, standardization, and matching values provided by Oracle Product Data Quality.

  3. Use the iterator control to examine different items in the set. Click Remove item for an item to be removed from the set.

  4. If the data quality results are satisfactory, click OK to apply the new values to the Create Multiple Items page.

  5. The data quality checks are also performed when you click Save and Close or Submit.

    Note

    You can submit items only when their item class is enabled for new item requests. If the data quality results contain a mixture of enabled and nonenabled items, then the Submit action is replaced by the following actions:

    • Add all of the items to a new item request, regardless of whether their item class is enabled.

    • Add only the enabled items to a new item request.

Editing an Item

To check data quality when you are editing an item interactively:

  1. On the Edit Item page, select Actions - Check Data Quality.

  2. Examine the Check and Apply Data Quality Results window, which displays the classification, standardization, and matching values provided for the item by Oracle Product Data Quality.

  3. If the data quality results are satisfactory, click Apply to apply the new values to the Edit Item page.

  4. The data quality checks are not performed when you click Save, Save and Close, or Submit.

    Note

    Data quality is not automatically checked when you update an item. To ensure data quality for existing items, you can run the periodically scheduled Semantic Key Update process.

Running a Scheduled Process

To check data quality for items in an item class by running a periodically scheduled process:

  1. On the Scheduled Processes Overview page, schedule the Semantic Key Update process.

  2. In the Process Details dialog box, select the periodic schedule for checking data quality.

  3. Select the parameters for the process:

    • Select the item class for which to check data quality.

    • Select Process All Items to check data quality for all items in Oracle Fusion Product Hub, not just those in the selected item class. By default, Process All Items is not selected.

  4. Submit the process.

Importing an Item Batch

You can specify data quality options when defining a source system or through an import batch definition by selecting Check Data Quality in the Data Quality section of the source system or batch definition. Items being created in Oracle Fusion Product Hub go through the data quality check automatically, while items that are updated using batches must have the data quality check initiated manually.

FAQs for Check Data Quality

When does data quality checking occur?

Data quality is checked automatically when you save newly created items, but you must explicitly check data quality when you save existing items that you have just updated. You can run a scheduled process to periodically check data quality.

When you are creating a new item, or multiple items, you can select Actions - Check Data Quality. The data quality checks are also performed automatically when you select Save, Save and Close, or Submit.

When you are editing an existing item, you must select Actions - Check Data Quality . Data quality is not checked automatically.

To ensure data quality for existing items, you can run the periodically scheduled Semantic Key Update process. This process checks data quality for all items in an item class. If any items are affected by classification, standardization, or matching, they are added to the Unconfirmed tab of a new item batch. You can then manage the unconfirmed items. After they are imported, the items are available with the changes caused by data quality checking.

When you are importing a batch of items, you can check data quality either automatically or manually. To enable automatic checking during data upload, select Check data quality in the Data Quality Options section of the Create Item Batch page. If automatic checking is not enabled, you can select a batch on the Manage Item Batches page then select Actions - Check Data Quality.

How can I use attribute groups to control data quality?

While defining an item class, select the attributes to be used when checking data quality in the order that you want them to be applied. All selected attributes are used for classification and standardization. You can also select attributes to be used for matching duplicate items.

To use attribute groups to control data quality:

  1. On the Edit Item Class page, navigate to the Attribute Groups and Pages tab.

  2. Select the Data Quality subtab.

  3. Select Actions - Select and Add.

  4. In the Select and Add: Data Quality Attributes window, search for one or more desired attribute groups.

  5. Select an attribute from the attribute group to be used when checking data quality. Repeat for other attributes that you want to use.

    Important

    The sequence in which you select the attributes determines the sequence in which they are used in checking data quality

  6. Select the Matching check box to select an attribute for use in matching.

How can I extract the metadata for the semantic model?

Select the Extract Data Quality Metadata action for an item class or catalog, or schedule the Data Quality Metadata Extract process.

The AutoBuild feature of Oracle Product Data Quality uses metadata for item classes and catalogs to construct the semantic model required for data quality checking. See the documentation for Oracle Product Data Quality for details about transferring the extracted metadata.

Note

Before you extract metadata, you must first set up the hierarchies for item classes and item catalogs.

To extract the metadata for a catalog or item class, select it on the Manage Catalogs or Manage Item Classes page, and select Actions - Extract Data Quality Metadata. This action schedules the Data Quality Metadata Extract process, which extracts the metadata into an external file, including the child hierarchy of the catalog or item class, and all of its attribute information. You can examine the results of the process on the Manage Scheduled Processes page.

You can also schedule the Data Quality Metadata Extract process directly on the Manage Scheduled Processes page.

  • In the Process Details dialog box, select either Item Catalog or Item Class as the extract type for the metadata.

  • For the selected extract type, choose the name of the catalog or item class from the Value list.

  • Submit the process and examine the results.

Can I bypass the data quality check while creating an item?

No. If data quality checking has been implemented, then the checks are performed automatically when you select Save, Save and Close, or Submit from the Actions menu. (The Submit action applies only to new item requests.)

What happens if I remove some items from the data quality results for multiple items?

Any items that you remove from the data quality results for multiple items are not created.

What happens if the data quality results for a new item are not satisfactory?

If you consider the data quality results for the item you are creating to be incorrect, then your immediate choice is to cancel the creation. However, you should also contact the administrator responsible for the setup of the semantic model and discuss the issue so that the model can be altered as required.

What happens if I do not accept the results after running a data quality check?

You must either accept all of the results of classification and standardization, or choose not to create the affected items. However, if new items are affected by the checks for matching, then you can ignore the duplicates and continue with creating the items.


Previous Page Next Page

Copyright © 2013, Oracle and/or its affiliates. All rights reserved. Legal Notices