Histograms can be used to estimate the selectivity of queries in query optimization. It is an unsolved problem using batched queries in compressed databases to construct an adaptive histogram to optimize query processing or answer queries approximately. This paper proposes to track hot data in compressed databases by scheduling these batched queries and use the feedback in query results to accelerate the convergence speed of the constructed adaptive histogram which can be maintained incrementally. A parametric method is also proposed to estimate the tuples falling in query area which is not covered by any bucket in the histogram. Experimental results show that the adaptive histogram has more average accuracy, higher convergence speed and better adaptability than STHoles.