As mentioned by MF_query, the IEFBR14 utility is normally used to delete the datasets in a job wch are to be used again. By setting the DISP ie the disposition parameter you can retain or delete or uncatlg the dataset.
1 MOD: the dataset already exists. If its to be created it ll be NEW instaed of MOD
2 UNCATLG: Uncatalog the dataset if the the job runs fine.
3 UNCATLG: Uncatalog the dataset if the the job does not run fine.
1 NEW: Dataset not on disk, to be created new in the further steps.
2 CATLG: Catalog it if the job runs fine.
3 DELETE: Delete it if the job does not run fine.