#!/usr/bin/perl

@words = qw (AS76573 
             AEDFG103020 
             A2D104380 
             S103282 
             ASD6D6383884 
             GH63737 
             AS76573AEDFG103020A2D104380S103282ASD6D6383884GH63737
             );

# RETRIEVE DATA AND APPLY MODS
$pkey = 0;
foreach $orig_word(@words){
  $pkey++;
  $new_word = $orig_word;
  $new_word =~ s/([a-zA-z][0-9][a-zA-Z]|[^\d])//g;
  print "\nWord $pkey = $orig_word -> $new_word\n";
}
# done --

http://database.ittoolbox.com/groups/technical-functional/sql-l/sql-query-to-remove-texts-and-numbers-5738855

[The original problem was stated as a SQL (SQL Server 2008) problem. This Perl script demonstrates the desired outcome for the problem statement.]

PROBLEM STATEMENT:

Please help me with SQL Query and requirement is as below: 

Below is the Data and I have to remove all texts associated with numbers. If incase in between LETTERS if we see number or numbers, those also should be removed, for ex:A2D104380, here A2D should be removed, others only letters. 

Field_Name 

AS76573 

AEDFG103020 

A2D104380 

S103282

To add to your point. 

Yes you are correct but gain if exists number and again followed by any 
letters then number also should be removed. 

For ex: ASD6D6383884 

Here ASD6D should be removed. 
GH63737. Here only GH should be removed 

SOLUTION RESULTS:

Word 1 = AS76573 -> 76573

Word 2 = AEDFG103020 -> 103020

Word 3 = A2D104380 -> 104380

Word 4 = S103282 -> 103282

Word 5 = ASD6D6383884 -> 6383884

Word 6 = GH63737 -> 63737

Word 7 = AS76573AEDFG103020A2D104380S103282ASD6D6383884GH63737 -> 76573103020104380103282638388463737

RegEx string substitution

regex, perl

Perl