I have a table into which users can enter a To and From range, and an application returns all the values in between. Naturally the users can enter to and from values that don't exist, as all they need to do is satisfy a "between" expression. eg they can enter '1100' to '1399' and even though neither value actually exists in the data, the query will return all the values that fall into the desired range.
However the application now demands a bit more of a sophisticated approach. The target data may have the form 'nn-nnnn-nn' where n is '0' to '9', and the user wants to select all the values (in pseudo-code) which satisfy a pattern like '%-7040-%' and '%-7400-%'
A simple LIKE expression with a mask containing a RegEx won't work - ie LIKE '%-7[0-4][0-4]0-%' doesn't work because (eg) '00-7150-00' won't satisfy the mask because of the 5 falling outside [0-4], although it is within the range I want.
I have come up with a solution to this problem by using MIN and MAX aggregate functions on LIKE [From] and LIKE [To] respectively and this works fine. eg
"SELECT A.<something> from <target_table1> A, <table_with_ranges> B WHERE A.<something> BETWEEN (SELECT MIN(X.<something>) FROM <target_table1> X WHERE X.<something> LIKE B.[From]) AND (SELECT MAX(X.<something>) FROM <target_table1> X WHERE X.<something> LIKE B.[To])"
It's actually quite a bit more complicated than that - I have simplified the SQL a bit, but that is the basic algorithm.
The problem is that for the BETWEEN expression to work, both the MAX and MIN functions have to return a value that is not NULL- ie there has to be a value in the target table which satisfies the mask. In the application, this is not always going to be the case - Users will want to create large ranges to allow for future growth so every time they enter new data into target_table1 the don't then have to go out and redfine the [From] and [To] ranges to satisfy the new data. What I really need is an SQL expression along the lines of "SELECT .... FROM ... WHERE A BETWEEN <pattern1> AND <pattern2>" where <pattern1> and <pattern2> contain wildcards.
Anyone got any suggestions??? By the way, using SUBSTRING or other string functions is NOT an option for a variety of reasons.
The most direct and efficient means of handling this is to parse your field into its three component parts. You can do this on the fly each time you execute your statement, but if you do this frequently and speed is important then consider breaking it into three columns in your database, or add three calculated columns the automatically parse it into its components.
If it's not practically useful, then it's practically useless.
Thanks blindman, - unfortunately the target table is selected on the fly (the application's purpose is to build a user-defined hierarchy over tables on any target system) so I have absolutely no knowledge of the target data, apart from type, so I can't do the obvious and use SUBSTRING or something similar.
I only picked the data in the original post as a typical example - not a definitive one.