Matching in MySQL : A comparison between REGEXP and LIKE

Abstract: When needing to search for data in multiple datasets there is a risk that not all da-tasets are of the same type. Some might be in XML-format; others might use a re-lational database. This could frighten developers from using two separate datasets to search for the data in, because of the fact that crafting different search methods for different datasets can be time consuming. One option that is greatly overlooked is the usage of regular expressions. If a search expression is created it can be used in a majority of database engines as a “WHERE” statement and also in other form of data sources such as XML. This option is however, at best, poorly documented and few tests have been made in how it performs against traditional search methods in databases such as “LIKE”. Multiple experiments comparing “LIKE” and “REGEXP” in MySQL have been performed for this paper. The results of these experiments show that the possible overhead by using regular expressions can be motivated when considering the gain of only using one search phrase over several data sources.