Sharing Data Prior to Publication. What Would You Do?

A researcher, speaking at a conference, presents two sets of results based on two related datasets. He references a paper that he published recently, describing in great detail the first set of results based on the first dataset. This paper mentions the second dataset but does not discuss the results based on it. A colleague attending the conference asks for a copy of both datasets. The researcher is pleased to deliver the first set of data, but does not want to turn over the second until he has had time to prepare another paper, describing results based on that dataset, and had it accepted for publication. Is the researcher justified in withholding the second dataset until he has published his results?