I have a 68 million row table with 5 integer columns and 1 column of text labels. What is the best way to index the text column? Cardinality is high (9.7 million distinct values). The default treatment creates a .sp file only and takes over 3 seconds to run (integer columns are < .01 seconds). Switching to datatype= keyword creates .dic and .idx files but takes over 20 seconds. Should I be using a .tdlist file?
my queries are of the form "select a,b,c where t='asdf123'" Andrew _______________________________________________ FastBit-users mailing list [email protected] https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users
