This table is built by a dedicated component during initial content processing. Assuming that the source content does not change, the table should remain valid as long as the document contains the same files.

The table is used during editing to implement the auto-propagation feature, whereby the translation of a repeated segment can be immediately propagated to other locations in the same document.

A RepetitionId is used to identify a set of source segments with the same content (repetitions) in the table. The actual segments that are associated with the repetition can be retrieved by calling GetRepetitions(RepetitionId). A segment is referenced as a paragraph unit id and the associated segment id (this is necessary since the same segment ids can be re-used in multiple files in the document (in fact, it is only guaranteed to be unique within a paragraph unit).

It is left to the implementation to determine how repetitions are calculated and propagated. Sophisticated algorithms can apply repetitions also when the source content differs, for example, with regards to tagging and placeables. Very often, a temporary translation memory is used to build the repetition table.

When a translation is applied using auto-propagation, the corresponding repetition id should be stored in the RepetitionTableId property. This allows the editor to better track auto-propagated segments.

Reference