Table search_backend_data too big

Hi! i have a legacy pimcore project with a very big table called search_backend_data

-rw-rw---- 1 mysql mysql 14G dic 12 09:33 search_backend_data.MYD
-rw-rw---- 1 mysql mysql 5,9G dic 12 09:33 search_backend_data.MYI

my question is: it’s used only in searches accross the interface? can i truncate it safely?

That is only used for backend search. So you could truncate it, but, it get’s re-indexed on save of every DataObject/Asset/Document again. To get rid of that, you could create a CompilerPass and remove the Listeners for the BackendIndex

reindexed completely? i mean, on the first saving process will regenerate the whole table or only the rows of the saved asset?

only for the single element