By Meta S. Brown

Data miners often sort cases (change the order of rows) to get clearer organization for viewing data or export. Or, you may have a functional reason to sort. For example, some applications require sorting data before merging (joining columns from different data sources).

The steps for sorting vary a lot from one application to another. These are common options that you may find in your data-mining application:

  • Interactive sorting in a data viewer: Look at the first variable, Clump.

    image0.jpg

    The values change from case to case, with no particular order. One click on the column header (where you see the variable name) sorts the data in ascending order of that variable.

    image1.jpg

    A second click re-sorts the cases in descending order. The Restore Order of Examples button returns the data to the original order.

    image2.jpg

  • Specialized sorting tools: Data-mining applications often have sorting tools. Look for them in menus for data manipulation or transform. In applications with many tools, you may need to use search to find the right tool; just search for sort.

    image3.jpg

    This icon represents the search function in a process. These tools can often handle complex nested sort schemes involving several variables.

    image4.jpg