Re: Indexing strategy

In article <u7l11ac.735908674@sun1> spitz@ana.med.uni-muenchen.de (Richard Spitz) writes: >Hi Informix Gurus, >..... >I have several large tables under ONLINE 4.0 (soon to be upgraded to 4.1), >whose structure is similar to: > > patient_nr INTEGER, > procedure_nr INTEGER, > proc_date DATE, > result SMALLFLOAT > >There can be only one procedure of the same type per day for one patient, >so the primary key is composed of patient_nr, procedure_nr and proc_date, >and I defined a composite unique index on those three columns. Since there >will be many queries on a per-patient, per-procedure or per-date basis, >I also defined duplicate indices on each one of these rows. > >..... > >The question is: Will INFORMIX be able to use the composite index to query >columns that are part of it with the same speed as with the additional >indices that consist only of the specific column? The composite index will be used for any query which includes the leading column(s) of the composite. Therefore, the duplicate index on patient_nr is redundant; however, the others will be required. >Thanks for any help, > >Richard ___ ___ Consultant, Client Srvcs Engineering / ) __ . __/ /_ ) _ _ __ Informix Software Inc. (303) 850-0210 _/__/ (_(_ (/ / (_(_ _/__> (-' ~/ '(_- 5299 DTC Blvd #740 Englewood CO 80111 {uunet|pyramid}!infmx!dberg #include<disclaimer.h>