Wednesday, September 20, 2017

REGEXP for searching in strings

regexe instr to find instances of character in search string

Continuing my posts about the regular expressions introduced as part of IBM i Technical Refreshes to 7.2, TR6, and 7.3, TR2, I have struggled to come up with a good brief description that would suffice for this post's title. IBM's documentation describes this function as:

returns the starting position or the position after the end of the matched substring

After playing with this function for awhile I would describe it as: "returns the position where the searched pattern occurs, whether first or nth occurrence, in the source string". And I would have to admit that description is better than IBM's, but it not a good description either. It is probably best I start explaining how to use this function and you will be able to decide how you would describe it.

The format for this function is as follows:

  REGEXP_INSTR(source-string, search-pattern, starting-position,
               occurrence,position-value, flags, group)

Some of the parameters' function is obvious, others less so.

  1. source-string  The variable or string, that will be searched.
  2. search-pattern  What I am looking for in the source string.
  3. starting-position  Where I want to start searching in my source string.
  4. occurrence  Do I want the first, second, third, etc. occurrence of the search pattern in the source string.
  5. position-value 
    • Zero ( 0 ), which is the default, return to me the starting position of the search pattern.
    • 1 return to me the ending position of the search pattern.
  6. flags  These are the same flags used for the REGEXP_LIKE.
  7. group  Capture group, I have to admit I did not find a reason why I would use this.

I am going to give examples of using this function in two ways:

  1. Using a Select statement, I would use this if I was using an SQL client, such as STRSQL.
  2. Using a Set statement, as I would if I was testing a string in a program.

Let's get started. Here is a Select statement containing the most basic form of the expression:

SELECT
  REGEXP_INSTR('RPGPGM.COM is the best website for rpg examples',
               'o',1,1)
FROM SYSIBM.SYSDUMMY1

In this statement I am searching the string for the lower case letter o only, starting in the first position, and for the first occurrence of a lower case o. As this is a Select statement it needs to be performed using a table/file, in this case I am using IBM's dummy table, SYSDUMMY1 in the library SYSIBM. The statement returns the following value:

REGEXP_INSTR
          33

What if I wanted to return the position of the first letter o of either case? I can use the lower case i in the flags perimeter, and I have to give a value in the position value too.

SELECT
  REGEXP_INSTR('RPGPGM.COM is the best website for rpg examples',
             'o',1,1,0,'i')
FROM SYSIBM.SYSDUMMY1

The position value, fifth parameter, is 0 (zero) as I want to return where the search pattern starts in the string.

This returns:

REGEXP_INSTR
           9

What happens when I change the position value to 1?

SELECT
  REGEXP_INSTR('RPGPGM.COM is the best website for rpg examples',
             'o',1,2,1,'i')
FROM SYSIBM.SYSDUMMY1

The value 10 is returned as that is where the current search pattern ends in the string.

REGEXP_INSTR
          10

And what about finding, for example, the second occurrence of the letter e in the search string?

I change the fourth parameter, occurrence, to 2.

SELECT
  REGEXP_INSTR('RPGPGM.COM is the best website for rpg examples',
             'e',1,2,0,'i')
FROM SYSIBM.SYSDUMMY1

I am returned 20, which, of course, is where in the search string the second e is found.

REGEXP_INSTR
          20

A better example of using the occurrence parameter is shown in the RPG example code below.

I create this program to show how I would do all of what I described above in a RPG program. Let me start with the definitions of the variables I will be using.

01  **free

02  dcl-s String char(47)
            inz('RPGPGM.COM is the best website for rpg examples') ;
              // ----+----1----+----2----+----3----+----4----+--
03  dcl-s Position int(5) ;
04  dcl-s Counter like(Position) ;

Line 1: This program is in totally free RPG, as it is 2017 why would I be writing in anything else?

Line 2: This is the definition of the variable to contain the search string I will be using. I created a ruler in a comment line below to make it easier to see where the various characters are.

Line 3: I will be using Position to contain the value returned by the SQL statements.

Line 4: Counter will be used to condition a For-loop.

Now onto the first example: returning the first occurrence of (lower case) o. My personal preference is to use SET, rather than SELECT, if I am not getting data from a table/file.

05  exec sql SET :Position = REGEXP_INSTR(:String,'o',1,1) ;
06  dsply ('1 = ' + %char(Position)) ;

The RPG variables used in the SQL statement need to be prefixed with :, so that the compiler knows this is a program variable.

When the program is run I see the following displayed:

  DSPLY  1 = 33

In the next example I wanted to find any letter o, regardless of case.

07  exec sql SET :Position = REGEXP_INSTR(:String,'o',1,1,0,'i') ;
08  dsply ('2 = ' + %char(Position)) ;

And I find the first one in the ninth position of the search string.

  DSPLY  2 = 9

In RPG I can illustrate better how to use the occurrence parameter to find multiple occurrences of a letter in the search string.

09 for Counter = 1 by 1 to 47 ;
10    exec sql SET :Position = REGEXP_INSTR(:String,'e',1,:Counter,0,'i') ;
11    dsply ('3 = ' + %char(Counter) + ':' + %char(Position)) ;

12    if (Position = 0) ;
13      leave ;
14    endif ;
15  endfor ;

Line 9: I am using the variable Counter as the counter for the For-loop. I am looping up to 47 as that is the length of the search string variable.

Line 10: The first time through Counter is 1, so it finds the first occurrence of e. Second time through I am looking for the second occurrence, etc..

Line 11: I am using the Display operation code to display the value of Counter followed by the position the e is found in.

Lines 12 – 14: If an occurrence of e cannot be found, then zero is returned from the SQL statement. As there are no occurrences to look for it is time to quit the For-loop.

When run I run this code I see:

  DSPLY  3 = 1:17
  DSPLY  3 = 2:20
  DSPLY  3 = 3:25
  DSPLY  3 = 4:30
  DSPLY  3 = 5:40
  DSPLY  3 = 6:46
  DSPLY  3 = 7:0

The seventh loop returned zero as there are only six e in the search string.

 

If you are interested in learning what else you can use the REGEXP_INSTR for visit IBM's KnowledgeCenter page, linked below, and use your favorite search engine to search for REGEXP_INSTR as much of what Oracle's version can do I have found the IBM i will do the same.

 

Other Db2 for i regular expressions:

 

You can learn more about the REGEXP_INSTR function from the IBM website here.

 

This article was written for IBM i 7.3 TR2 and 7.2 TR6 only.

6 comments:

  1. Great info I can see how this might help with certain parsing routines.
    *FYI line-9 2nd to last word is STRONG seems like you meant STRING

    ReplyDelete
  2. Thank you for bringing the spelling error to my attention. It has now been fixed.

    ReplyDelete
  3. Power i and modern powerful pattern searching with regular expressions. Know the tools you have so you will know when to use them and how.

    ReplyDelete
  4. Nice article. Thanks.

    Just a note.
    Instead of using SYSDUMMY1, we can use the command VALUES.
    i.e: VALUES(REGEXP_INSTR('RPGPGM.COM is the best website for rpg examples','o',1,2,1,'i'));

    ReplyDelete
  5. You always have helpful posts, whether as review or learning something new. I appreciate it. It's nice to see a very powerful Unix function in SQL. One correction to the above:

    "The position value, fourth parameter, is 0 (zero)". I think you meant the fifth parameter.

    ReplyDelete
    Replies
    1. Oops (hangs head in shame).
      Thank you for bringing that error to my attention. It has now been fixed.

      Delete

To prevent "comment spam" all comments are moderated.
Learn about this website's comments policy here.

Some people have reported that they cannot post a comment using certain computers and browsers. If this is you feel free to use the Contact Form to send me the comment and I will post it for you, please include the title of the post so I know which one to post the comment to.