The problem
Real data holds the same thing spelled several ways: Houston, houston, Huston, Houston, TX. Grouped incorrectly, they count as four cities.
Profile-pane grouping fixes this without a lookup table.
Manual grouping
Select several values in the profile pane, right-click, Group. They are merged under one label, which you can rename. Prep records it as a change you can edit later.
Fine for a handful of values. Beyond that, use automatic grouping.
Automatic grouping
Field menu → Group Values, then choose a method:
| Method | Matches on |
|---|---|
| Pronunciation | Values that sound alike — good for names |
| Common Characters | Same characters in any order — catches transpositions and punctuation differences |
| Spelling | Small edit distance — catches typos |
| Manual selection | You choose |
Pronunciation handles Smith / Smyth. Common Characters handles Houston, TX / TX Houston. Spelling handles Huston.
Always review
Automatic grouping will confidently merge things that are genuinely different — two real customers with similar names, or two product codes differing by one character that matters.
Prep shows you the proposed groups before applying. Read them.
The risk is asymmetric: an ungrouped duplicate is visible in the output, whereas an incorrectly merged pair silently disappears into one row.
When to fix it upstream
If the same inconsistency arrives every month, grouping in Prep fixes the symptom repeatedly. A validated dropdown at the point of entry, or a mapping table joined in the flow, fixes it once.