Log files are an important by product of any computing systems including database systems. They contain a huge amount of historical data. Although many algorithms have been designed to utilize the information stored in such files, many of them can still be further improved in terms of execution time and memory usage. In this research paper, a binary-based approach for mining frequency of data items in database transaction log files is introduced. Both the data structures and the algorithms used will be presented according to the sequence of the methodology stages carried out in this research work. The stages are pre, during and post scanning of the log file. Initial experimentation of the approach reveals a significant improvement in terms of the execution time taken to perform frequency analysis of a database transaction log file. To validate the approach, performance comparison was also done against the popular Apriori algorithm. Initial result has shown enhancement in terms of execution time using the binary-based approach. © Springer Science+Business Media Singapore 2014.