Patient attributes are counted at sample level in combined study when one study lacks attribute

[alisman]

genie:

http://localhost:8082/study/summary?id=crc_public_genie_bpc%2Cacc_tcga_pan_can_atlas_2018

The attribute AJCC_STAGING_EDITION is being counted as NA for all SAMPLES. It is a patient level attribute belonging only to one of the studies in this combined study view. It would seem the legacy is doing the right thing counting it patient level.