The applicant was using two governmental data bases, ‘Current Status of Information Processing’ survey by METI and ‘Economic Census’ by the MIAC for his research.  As the data structure of these databases, as well as their field names are not consistent, collating firm records in these two databases were a time-consuming process.  Due to their confidentiality, it was not possible to outsource recompilation, involving some 500 Excel tables, forcing the applicant to do the work himself manually.  This project aimed at automating the process of recompilation, or at least simplifying the manual recompilation.  The main difficulty was that the codes allocated to sample firms were totally different between the two databases.  Using the firm name was problematic as there were number of firms with the same names.  Combining the firm name and the address was not satisfactory either because the way address was written was not unique, requiring human judgement.  Overall, efficiency gain of about 50% was achieved in term of the process time, by improving the manual process with some automation.  The summary of the project and its result and discussions are planned to be reported at the September meeting of JASMIN.