From: gustavo <gu...@mw...> - 2004-05-20 17:36:01
|
Hi, Is there any plans to integrate Yaffil optimizations in a short = timeframe? I=B4m asking because some benchmarks I=B4ve made using some big tables = (300K recs) shows that Yaffil always outperforms FB 1.5: create table mov_pe(client int,shipper int,salesperson int, = delivery_date date, product int, color int, pattern int, size varchar(5), qty numeric(15,7), price numeric(15,2),mov_type int); Create index idx_1 on mov_pe(mov_type); Create index idx_2 on mov_pe(product); Create table mov_types(mov_type int, mov_code varchar(1)); Create unique index idx_3 on mov_types(mov_type); Create table products(product int,brand int); Create unique index idx_4 on products(product); Create index idx_5 on products(brand); Create table brands(brand int,brand_name varchar(50)); Create unique index idx_6 on brands(brand); Avg table sizes: mov_pe: 300.000 recs mov_types: 30 recs products: 6.000 recs brands: 15 recs The query is: Select brand_name,sum(qty),sum(qty*price) from Brands b Inner join products p on b.brand=3Dp.brand Inner join mov_pe mp on mp.product=3Dp.product Inner join mov_types mtp on mtp.mov_type=3Dmov_pe.mov_type Where mtp.mov_code=3D'V'; Execution time for this query: FB 1.3: 9 secs FB 1.5: 7.5 secs Yaffill 872b 4 secs As you see its a huge difference. FB 1.5 is a lot faster than FB 1.3, = but merging Yaffil optimizations can make FB 1.? Faster even faster than = Yaffil. Thanks for your attention and for your great contribution, Gustavo Hispagnol Millennium Network --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.688 / Virus Database: 449 - Release Date: 18/05/2004 =20 |