We briefly review some recent work on hypercircle inequality for partially corrupted data when the data error is measured with l∞ norm. The aim of this paper is to present the method for approximate missing data in the use of midpoint algorithm and discuss some numerical experiments in the hardy space which is well known in a reproducing kernel Hilbert space (RKHS).