Performance improvement of join algorithms for efficient query optimization with early and maximized result rate

Abstract

Enhancing the performance of large database systems depends heavily on the cost of performing join operations Optimizing the join operation of two very large tables is considered as one of the interesting research topics to many researchers especially when both tables to be joined are very large to fit in the main memory In such cases join is usually performed using hashing or sorting technique In hash based join the database needs to hash and partition the build input and with a sort merge join the database needs to sort the input before it can produce the join results However both hashing and sorting are considered as blocking operations since they block the progress of join operation until they are completed In recent years there are a large number of join queries which are being executed by the interactive users and applications In all the interactive applications the time to produce the first few results are very crucial The state of art join algorithms are not ideal for this setting as most of the algorithms are hash sort based algorithms which require some pre work before it can produce the join results Hence four new join algorithms are proposed in the present research which can produce join results at higher rates during the early stages of the join operation These proposed join algorithms are implemented without hashing and sorting techniques newlineThe proposed MRR join algorithm can be used to produce early and maximized join results during the earlier stages of the join operation This is achieved by exploiting the distribution of the data in the join attribute column which is available in the histogram newline newline