Values analysis is often useful as the first type of analysis to perform on data which is relatively unknown to the analyst. It helps determine the nature and overall quality of the data. For example, whether the data is categorical or continuously numeric, how many null values it contains, and so forth. Values analysis can readily be applied to any type of character or numeric data, even date fields.
Given a table name and the name of a column, the Values analysis provides a count of the number of rows, rows with non-null values, rows with null values, rows with value 0, rows with a positive value, rows with a negative value, and the number of rows containing blanks in the given column. Optionally, unique values are calculated within the analysis as well.
Note that for a column of nonnumeric type, the zero, positive and negative counts will always be zero (for example, “000” is not counted as 0).
If multiple columns are requested, a VOLATILE table is built, and all columns are processed in a single CREATE VOLATILE TABLE AS SELECT… statement. Data is reformatted with individual INSERT/SELECT statements into the final output dataset as described below. In this case, the create view option may not be requested.
The Values analysis is parameterized by specifying the table and column(s) to analyze, options unique to the Values analysis, as well as specifying the desired results and SQL or Expert Options.