Fragmenting Indexes

Version: IDS 10 I am currently looking at various fragmentation strategies for Indexes that have keys of type CHAR. I have following questions: 1. Columns that have value of "000987" but is a CHAR, which strategy makes more sense, doing the fragmentation by using SUBSTR or using MOD of HASH. Is there any such HASH function in IDS that could be used for such scenarios. Is there any other fragmentation strategy that is better than above. 2. Column that are CHAR and are typically string, what would be a better fragmentation strategy. What options do I have other than SUBSTR. 3. Is substr efficient ? Does it slow down building indexes