Yes, he may, provided the data is reliable and adequate for the purpose of hazard classification. Further to this, available studies should be sufficiently documented to assess their quality and adequacy.
The physical hazards of substances and mixtures should be determined through testing based on the methods or standards referred to in part 2 of Annex I to CLP. These methods can be found for example in the UN Manual of Tests and Criteria, which is normally used to classify substances and mixtures for transport:http://www.unece.org/trans/danger/publi/manual/manual_e.html
However, testing is not mandatory in cases where adequate and reliable information from reference literature or databases is already available and where the substance to be classified and the substance described in the reference are comparable with regard to homogeneity, impurities, particle size etc.
Open literature or databases often use secondary data sources. When such data is used, the original source should be cited and checked by an expert. This should involve the check that there is sufficient documentation to assess the suitability of the test used, and that the test was carried out using an acceptable level of quality assurance. Useful data compilations containing physicochemical data are listed in section R.220.127.116.11- 'Available information on physicochemical properties' of the Guidance on information requirements and chemical safety assessment on ECHA's website: http://www.echa.europa.eu/guidance-documents/guidance-on-clp