SQL Data Extraction with Regular Expressions

Regular expressions (RegEx) are a powerful mechanism for matching patterns in text, allowing us to extract specific characters, validate input, and search for patterns. In SQL, regular expressions are used in queries to validate data, extract specific substrings, or search for matching patterns in complex strings.

This article will demonstrate how to use regular expressions in SQL through various scenarios such as validating email addresses, extracting alphanumeric values, and cleaning up data.

Examples of SQL Data Extraction with Regular Expressions

To demonstrate the use of regular expressions in SQL, we will be using an existing sample table named Country under the GEEKSFORGEEKS database. This table contains records with country names, which will serve as our data for applying regular expressions.

Example 1: Validating Country Names Using LIKE and Regular Expressions

1. Match Country Names Starting with Letters A - D and the Second Letter Between U - Z.

In this example, we use the LIKE operator with a regular expression to filter country names. The query matches country names where the first letter is between A - D, and the second letter is between U - Z, followed by any characters.

Query:

SELECT * FROM Country
WHERE CountryName like '[A-D][U-Z]%' 

Output

2. Match Country Names Starting with Letters A, B, C, or D, and Any Subsequent Characters

If we need to specifically check for the first character alone and the rest of the characters can be anything, then 

Query:

SELECT * FROM Country
WHERE CountryName like '[A-D]%' 

Output

3. Match Country Names Starting with 'U'

If we need to filter country names that specifically start with the letter 'U', we can write a simpler query.

Query:

SELECT * FROM Country
WHERE CountryName like 'U%'  

Output

Suppose if we want to find country names starting with 'U'  and additional information then the query will be as follows:

Query:

SELECT * FROM Country
WHERE CountryName like 'U% [S]%'

Output

Example 2: Using PATINDEX to Find Specific Patterns

1: Find the Position of Numeric Characters in a String

The PATINDEX function can be used to locate the position of numeric characters within a string using a regular expression. By specifying a pattern that searches for non-alphabetic characters (e.g., numbers), PATINDEX returns the position where the first numeric character appears in the string

Query:

SELECT 'GFGVersion1' as InputString,
PATINDEX('%[^A-Za-z]%', 'GFGVersion1') as
NumericCharacterPosition;

Output

2: Find the Position of the First Numeric Character

To get numbers only from an input string, we can use in below way also

Query:

SELECT 'GFGVersion1' as InputString, 
PATINDEX('%[0-9]%', 'GFGVersion1') as
NumericCharacterPosition;

i.e. instead of using [^A-Za-z], used [0-9] and getting the same result as above

-- 0 will indicate no numeric value present
SELECT 'VERSION' as InputString, 
PATINDEX('%[^A-Za-z]%', 'VERSION')
 as NumericPosition; 

Output

Example 3: Removing Numbers from a String Using STUFF

We can use this functionality of using a regular expression with functions like PATINDEX, we can solve to get only the characters alone from input string/numbers alone from input string etc.,

Query:

SELECT STUFF('GE098EKS9VER1', 3, 1, '' );    

Output

Remove All Numeric Values from a String

By having PATINDEX and STUFF functions, we can get only the character values from an input string.

• We need to use a regular expression to get applied on PATINDEX
• Find out the numeric position and remove the numbers by using STUFF function
• Step 2 has to be repeated until there is no numeric value

Query:

DECLARE @inputData NVARCHAR(MAX) = 'GE098EKS9VER1'       
--input string
DECLARE @intPosition INT    
 --get the position of the integer from the input string   
SET @intPosition = PATINDEX('%[^A-Za-z]%', @inputData)    
print @intPosition

--run loop until no integer is found in the input string
WHILE @intPosition > 0                                     
  BEGIN  
     --remove the integer from that position
    SET @inputData = STUFF(@inputData, @intPosition, 1, '' )   
    --PRINT @inputData 
    SET @intPosition = PATINDEX('%[^A-Za-z]%', @inputData )
    --Again get the position of the next integer in the input string
    --PRINT @intPosition 
END  
SELECT 'GE098EKS9VER1' as InputString, @inputData AS AlphabetsOnly

Output

Conclusion

Regular expressions are a flexible tool in SQL that enable pattern matching, validation, and data extraction. By using functions like LIKE, PATINDEX, and STUFF, we can easily validate complex data such as email addresses, extract numeric or alphabetic characters from strings, and clean up data. Understanding and applying regular expressions in SQL queries will significantly improve data validation and manipulation in our applications.