Dataset Viewer: Exception Detection

Dataset View is a flexible tool that can help the User to detect anomalies and exceptions in the Datasets. This article describes Dataset Viewer settings in detail.

Table of contents:

  1. Fields Selection
  2. Adding Derived Fields to Dataset's Result Set
  3. Defining Filters
  4. Types of Filters
  5. AND/OR Operators
  6. Comparing Dataset Date Fields to Snapshot (Historical) Dates
  7. Tracking Changes
  8. Saving Field Selection and Applied Filters as a Separate View

1. Fields Selection

Fields Selector

You can limit the display of Dataset's All Data only to those fields that are required for your data analysis.

Just select the fields you need in the Select Fields section, click [Apply Changes] in the right side of the screen and the Result set is going to be updated and only those columns that you have selected are going to be displayed.

2. Adding Derived Fields to Dataset's Result Set

Derived fields include values that do not exist in a Data Source itself but are calculated from one or more existing numeric fields via basic arithmetic expressions and non-aggregate numeric functions.

For more details refer to:

3. Defining Filters

In addition to selecting specific fields, you can filter data using custom conditions. The process of creating a filter is similar to constructing a sentence: you choose a subject, comparison conditions and the values which are going to serve as a reference for filtering.

Defining Filters
Part Description Example
Subject The column you want to base the filter on. This drop-down list includes the same values as those displayed in the the Filed Selector to the left of the Filter section. name, units, country
Condition Contains a set of comparison operators which define the Subject. This drop-down list includes different values depending on whether the values in the selected subject are text or integer:

For text columns: exactly matches, contains, is in list, is not empty, does not equal, starts with.
For integer columns: equals, is more than, is more than or equal to, is less than, is less than or equal to, does not equal.

Some of the conditions define the relationship between the Subject and the Value that follows; others function independently and do not require a specific Value.
Different for text and integer fields
Value The Value of the subject you're basing the filter on: keyword, text or numeric value. Value may be added with wildcard (%%)representing alphanumeric characters. corporate sales, 5000, Spain

4. Types of Filters

You can use filters to seed out the required information using conditional filtering. You may mix and match several simple Rules or compound filters (Groups) as defined in Create a Dataset View  aggregated via [AND/OR] operators to focus on a highly specific data slice. The Results set will be displayed right on the same page.

Rule (Simple Filter Criteria)

Rule (Simple Filter Criteria)

Group (Compound Filter Criteria)

Group (Compound Filter Criteria)

5. AND/OR Operators

When you're using multiple Rules and Groups of conditions, you can use an AND/OR switches. You can not mix and match AND/OR for the set of applied filters but if you add a group of conditions, you may apply an alternative operator to it.

In the example below the following filters are applied to a Dataset. All of them are aggregated by the AND operator, meaning that all of 3 filters should be TRUE:

  1. Simple Rule
  2. A Group of 3 conditions aggregated by the OR operator (ANY of 3 conditions should be TRUE)
  3. A Group of 2 conditions aggregated by the OR operator (ANY of 2 conditions should be TRUE)
AND / OR Operators
Operator Description
AND We return all rows that match ALL of the filters.
OR We return all rows that match ANY of the filters.

6. Comparing Dataset Date Fields to Snapshot (Historical) Dates

Snapshot date is the date when the Dataset has been collected or updated.

Datasets may be:

  • Single Instance: Shown only when the Dataset has been collected (as shown in the picture below)
  • Multi-Instance (Snapshot dataset): Keeps history of Data updates

User can filter on a date field by comparing the field to the Dataset effective date for a single instance or multi-instance Dataset.

Multi-Instance Dataset Example (click to open)

7. Tracking Changes

You can compare the data from the current period to the prior period to highlight changes. For more details refer to this article: Snapshot Datasets: Comparing Instances.

8. Saving Field Selection and Applied Filters as a Separate View

There is no need to apply the same filters to review the same Results set later. After setting the Filters the User can save the Results set with selected fields and applied filters as a separate View by clicking [Save As View]. To learn more about Views, refer to: Create a Dataset View