Task Types
The type of task that a workflow assigns to you depends on the type of data quality issues that the source database tables contain. The records in the tables might contain errors, null values, or values that are inaccurate in the current data project. The tables might contain records that are redundant because they contain different versions of the same information.
You can work on the following types of task instance in the Analyst tool:
- Correct exceptions task
- Contains records that might include errors or null values. Analyze and fix any error that you find. When a record is free of errors, update the record status to indicate that the record is valid.
- Correct duplicates task
- Contains records that might contain duplicate information. The task sorts the records into clusters. A cluster is a group of records that represent the same business entity in the source data set. Analyze each cluster, and define a preferred version of the record that the cluster represents. Update the cluster status to indicate that you reviewed the preferred record.
- If a record is not a duplicate of any record in the cluster, move the record to another cluster. You can create a cluster that contains a single unique record.
- Review exceptions task
- Contains a set of exception records that an earlier user analyzed in a correct exceptions task. Review the work done by the earlier user, and verify that the record data and the record status are correct.
- Review duplicates task
- Contains a set of clusters that an earlier user worked on in a correct duplicates task. Review the work done by the earlier user, and verify that the preferred record data and the cluster status are correct.
Note: Records cannot pass from a task that corrects or reviews exception records to a task that corrects or reviews duplicate records. The database tables that contain exception records and duplicate records have different structures.