Table of Contents


  1. Preface
  2. Data Profiling
  3. Profiles
  4. Profile results
  5. Tuning data profiling task performance
  6. Troubleshooting

Data Profiling

Data Profiling

Data profiling tasks

Data profiling

You can create
data profiling
tasks in
Data Profiling
. Create and run a
data profiling
task to determine the characteristics of columns in a source object, such as value frequency, patterns, and data types.
Data profiling
tasks are also called profiles.
You can create a profile for a source object after you create a connection to the data source. After you create the profile, run it to view the profile results.
You can add a filter to run the profile on filtered results. For example, to view county-specific sales, you can filter the Sales table based on a county and then run the profile. You can add a schedule to run the profile at regular intervals. You can add rule specification, cleanse, and verifier assets as rules to a profile. You have to create these assets in
Data Quality
When you run a profile on a source object, the results include the following column statistics:
  • Number of distinct, non-distinct, and null values
  • Percentage of distinct, non-distinct, null, zero, and blank values
  • Documented and inferred data types
  • Number of patterns
  • Percentage of top pattern
  • Maximum and minimum length of values
  • Maximum and minimum values
  • Average, sum, and standard deviation for numeric data types
  • Value frequencies
  • Outliers
After you run a profile, you can perform the following actions:
  • Drill down on value, data type, and pattern to view drilldown results.
  • View historical and latest profile results
  • Create and run queries to view source rows that have data quality issues.
  • Compare multiple columns in a profile run
  • Compare two profile runs to analyze the statistics
  • Export profile results to a Microsoft Excel file
  • Monitor profile jobs in
    Data Profiling
    , or
    Operational Insights


We’d like to hear from you!