Analyzing Results
Match Rule Tab
This tab is divided into two sections:- The left pane displays the match rule. On expanding the Rule, you can view all the conditions and sub-conditions.
- A preview of these conditions is displayed in a tabular format on the right
pane of the screen. It shows these details:
- An Attribute such as Threshold, Scoring Method, Algorithms, and Missing Data Method.
- Value for each of these attributes.
Note: The Spectrum Smart Data Quality (SDQ) is integrated with Data Stewardship,
which helps you improve the match rules based on the exception handling done
in Data Stewardship. When you save the manual updates to the records in Data
Stewardship, it reflects as a notification on the
Projects page in SDQ, corresponding to the
project you made modifications to.
Note: Finally, the Data Stewardship Data Quality page provides information
regarding trends across data flows and stages.
Match Key Tab
This tab displays potential match key components in a tabular format. It also
displays the Column in which the match key component was
detected along with the Algorithm to be used. The
Average Group Size helps you determine the average
size of the group for your match key, which is generated for the complete
dataset. You get the average of each value based on each match key combination
to avoid any loss of accuracy. You can review and choose to consume any of the
potential match key components based on your scenario by adding these in the
Match Key Generator stage of Enterprise
Designer.
Note: As of now, only the Substring
algorithm is supported.
Example: This table displays a potential match key-
Match Key 1 detected in the
phone column with an average group size of 2. The
algorithm to be used is SUBSTRING (1, 7), where
1 is the starting index, and 7 is the last index
to be specified in the options of the Match Key Generator
stage. The starting index is fixed to 1 for all potential match key
components.
Match Key | Column | Algorithm | Average Group Size |
---|---|---|---|
Match Key 1 | phone | SUBSTRING (1, 7) | 2 |
Based on the actions performed by you : Variations present in the sample data uploaded, Columns selected for matching , and Records Tagged, the system has unlocked patterns present in your data to provide you with a match rule and potential match key components. It is suggested to test the generated results on your dataset.