When you load data into a data source, Discovery Center automatically calculates two types of statistics to represent your data:
-
Inferred. Derived from the full volume of data in the attribute.
-
Target. Required to hold all existing values in the current (target) attribute.
Each data source attribute has a target and inferred data type of either integer, decimal, or string depending on the structure of the data.
You examine data types for a single attribute by drilling down from an attribute's Structure > Data Type pie chart. You can also see the target and inferred data types for all attributes in a data source from the Attribute Summary tab.
To drill down from an attribute's Structure > Data Type pie chart
-
Open a data source.
-
Click Attribute Details.
-
In the Attribute Name list, select the attribute whose data types you want to examine. The Structure > Data Type pie chart graphically shows the distribution of data types in the attribute. The chart's key lists the data types found and their distribution percentage in the attribute. An attribute's data may consist of a single data type or a combination of types. If there are no values found in some rows of the attribute, the percentage of these rows shows as Null.
- Example:
-
This is the pie chart for the attribute Phone and shows both String and Integer data types in the attribute, with String the predominant type distributed across 88.55 percent of the attribute. Integer data represents 6.74 percent of the data, and 4.71 percent of the rows in the attribute contain null (empty) values.
-
Hover over the chart sections to see the value count for the data type. Nulls display the row count.
-
Double-click a section of the chart. A tab opens showing the attribute values for the data type selected. If you click a Null section, a tab opens showing the data rows that contain the empty data values for the current attribute.
-
To open a tab that shows the data rows that contain the string, integer, or decimal
values, double-click a row on the Strings: attribute_name, Decimals:
attribute_name, or Integers: attribute_name tab.