merged_java.csv contains an extended version of the Java dataset. The manually classified Java comments present in "Classifying code comments in Java open-source software systems" have been merged with the NLBSE'23 Tool Competition Java dataset. A conservative approach has been adopted to remove duplicates during the merging process, in particular, given only an overlap in files of 10%. Those files have been removed from the newly considered dataset. This approach also leaves the original NLBSE'23 Tool Competition Java dataset untouched.