Data Comparison Algorithm

While comparing, when 'Automatic sort' (see Set-up Data Comparison options) option is selected the program tries to get the info from the database where there are columns that are defined as Unique, i.e. contain unique values for each row of the table (indexed columns with no duplications, primary keys etc.). If such info is received the program assumes that the values in such columns are integrated row identifiers. The program sorts the data by these columns before comparing and aligns the rows by these identifiers while comparing:

if one of the compared recordsets has no record with such identifier, the program assumes that the record is missed;
otherwise the program compares the record values (values of the remaining columns) with the values of the first found record having the same identifier. If the values are different - the program marks the records as different.

If any "unique" column is not found, the program does not sort the compared recordsets and compares the first row of the left recordset with the first row of the right recordset etc.
User can manually specify the column that will be assumed as the row identifier. In order to do this:

In the connection tree right-click on the column by which you want to sort the data.
In the popup menu select Sort by this Column option.

When such "sort" columns are specified, the program will not try to get the info about unique columns and will assume the "sort" columns as the integrated row identifier. User can disable the program trials to sort the data automatically. In order to do this:

Select Tools->Comparison Options item of main menu. Comparison Options dialog is displayed.
Select Data Comparison tab.
Uncheck Automatic sort option at the bottom of list
Click OK button to save the settings.