It is possible to split segments in the Smartcat editor:
This can be useful if for example the segmentation was not done properly because of missing spaces before a period. Or in cases of a long sentence with many clauses where the translation would be written with multiple sentences that could re-used in the TM individually.
This process doesn't affect the source document - only the text in the Smartcat editor. Keep in mind that if this document is updated and the long sentence is encountered again in the future, the TM engine might not find a match if it was split in previous versions. It would have to be split again in the editor for Smartcat to offer more accurate matches.