NLM-CHEMSORT: an algorithm and computer program for sorting chemical names.

An algorithm is described that has been designed to sort medium-size lists of chemical names, including common, generic, trivial, and systematic names and code numbers, into a logical sequence. It successfully sorted more than 99.5% of 3767 names in its first application. Minor revisions then resulted in more than 99.9% success with the same set of names. The algorithm generates an 80-character primary sort key (alphabetic characters only) and a 16-character secondary level sort key (alphanumeric characters). These sort keys are generated de novo from the name as needed and, thus, do not require increased permanent-storage costs. Sorting on the primary sort key (and secondary sort keys when identical primary keys exist) results in logical sequences of chemical names.