Sometimes, for yet unknown reasons, inside the translation corpora there may be user translated segments with ids that do not exist in the corresponding source subsection/paragraph.
While it is not yet known, why this can happen, this issue blocks such draft translations from being continued, and we should change the application behaviour to silently fail to restore that specific segment/sentence, instead of leading to an error, and blocking the draft translation continuation.
The frequency of this issue is also unknown at this point.