Sometimes a picture helps. I’m working on dataset kept in Tinderbox. It tracks the ACM Hypertext Conference’s output and citations within-the conference corpus. So I update theTBX and public dataset once a year. Drawing on the publisher’s data (which invariably has errors) the dataset is hand-compiled. I’d just finished this year’s data and the publisher released the papers to its website using sequential numbering (a legacy hang-over from print: ACM no longer offers paper) instead of the per-paper the Conference had asked. Changing '2:1–8" to ‘10–18’, etc is a pain as there isn’t an easy source of which to script so manual edit (only 61 items is easier. Using Outline+column view makes it easy to spot re-keying errors.
Here, I’m editing in the text pane but watching the (second) column for sequence errors:
A good tip, especially in large/complex documents, is to turn off the columns when not needed. As long as that tab remains in the TBX, it will remember its last-used column settings (note: so the data is stored in the tab\s data, not discrete from it).
Another approach might be to use Attribute Browser view, but in this case Outline is easier for the task.