RPG Academy: BIF Up Your Code! Building Excel's PROPER Case Function and Testing Chg_Case

  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times

The previous TechTip focused on two of Excel's text conversion functions, UPPER and LOWER. This time around, I'll cover PROPER, also known as Title Case. Read on!


PROPER Case is a bit complex to implement, because it completely modifies the input string, changing the first letter of each word to uppercase and all of the others to lowercase. It's a perfect opportunity to show how a fairly complex problem can be solved in a simple way, using BIFs.


PROPER Case is a bit trickier to code than the functions presented in the previous TechTip, but if you think for a bit about the definition, you'll see that the answer is right there. We just need to find a way to pinpoint the beginning of each word of the string. This TechTip is a follow-up from the previous one in this series, so if you haven't read it, now it would be a good time to do so.


We're going to modify the Chg_Case function in order to increase its functionality. The first change is quite simple: we want to add a new case, so the validation of P_Case must be changed accordingly (I've added the line in bold):


* P_Case                                    

C                  IF       P_Case = ''    

C                             AND P_Case <> 'U'

C                             AND P_Case <> 'L'

C                             AND P_Case <> 'S'

C                             AND P_Case <> 'P'

C                   RETURN    P_String        

C                   ENDIF                    


The next step is adding a new WHEN statement to the SELECT that processes the input string. But before I do that, let's take a moment to analyze our problem: how to know which characters to capitalize. Just find the beginning of every word, right? One way to find out where each word begins is by looking for the blank space that separates it from the previous word. However, this approach doesn't work with the first word because we're trimming the leading blanks, so we need to treat it separately. The first word can be processed using the Sentence case approach from the previous TechTip. The following words of the string, however, require a little more processing:


* Proper Case (First Letter Of Every Word In Upper Case,                

*             The Rest In Lower Case) aka Title Case                    

C                   WHEN     P_Case = 'P'                                

* Process the first word                                                

C                   EVAL     W_TempStr = %XLATE(lc : UC :              

C                                               %SUBST(W_String : 1 : 1))

C                                         +                              

C                                         %XLATE(UC : lc :              

C                                               %SUBST(W_String : 2))  

C                   EVAL     W_Pos = %SCAN(' ' : W_TempStr)            

C                   DOW       W_Pos <> *Zeros                  

* Process the following words                                            

C                   EVAL     W_TempStr = %SUBST(W_TempStr : 1 :          

C                                                W_Pos) +                

C                                         %XLATE(lc : UC :              

C                                                 %SUBST(W_TempStr :      

C                                                       W_Pos + 1 : 1))  

C                                         +                              

C                                         %SUBST(W_TempStr : W_Pos + 2)  

C                   EVAL     W_OldPos = W_Pos                            

C                   EVAL     W_Pos = %SCAN(' ' : W_TempStr : W_OldPos + 1)

C                   ENDDO

C                   EVAL     W_Return = %Trim(W_TempStr)                                            


We're going to use %SCAN to distinguish between words. In this case, we're going to store the result of the %SCAN in the W_Pos variable. So what we're going to do is a three-step operation: first, we need to keep what was already convertedin other words, a substring starting with the first character of our string and ending with the blank space we've just found: %SUBST(W_TempStr : 1 : W_Pos); then we' re going to convert the first character after the blank space to uppercase: %XLATE(lc : UC : %SUBST(W_TempStr : W_Pos + 1 : 1)); for the third and final step, we're going to concatenate the rest of the string without making changes: %SUBST(W_TempStr : W_Pos + 2).


Next, we need to save the position of the last blank space found in W_OldPos and look for the next blank space, using the third parameter of %SCAN to start after that last blank space, thus making sure that we're not finding the same character twice. Since %SCAN returns zero when it can't find what it's looking for, we're using that to do the loop. Finally, we just have to pass our work variable to the one that's going to be returned (W_Return), and that's it.


Now, let's test our function! I've written a simple test program that tests the different Chg_Case variations:



*   Variables                                                           *


D W_MyString2   S           250A   VARYING INZ(*Blanks)                          

D W_TestStr2     S           250A   VARYING INZ(*Blanks)                



*   Copy Statements                                                     *


* String Operations                                                      

/Copy QCPYLESRC,STR_OPS_PR                                              


C                   EVAL     W_MyString2 = 'THIS is A Test'              

* Test the Chg_Case function (UPPER CASE)                          

C                   EVAL     W_TestStr2 = Chg_Case(W_MyString2 : 'U')

* "W_TestStr2" is now 'THIS IS A TEST'                            

* Test the Chg_Case function (lower case)                          

C                 EVAL     W_TestStr2 = Chg_Case(W_MyString2 : 'L')

* "W_TestStr2" is now 'this is a test'                            

* Test the Chg_Case function (Sentence case)                        

C                   EVAL     W_TestStr2 = Chg_Case(W_MyString2 : 'S')

* "W_TestStr2" is now 'This is a test'                            

* Test the Chg_Case function (Proper Case)                          

C                   EVAL     W_TestStr2 = Chg_Case(W_MyString2 : 'P')

* "W_TestStr2" is now 'This Is A Test'                            


C                   EVAL     *INLR = *On                            


I've created a service program called STR_OPS (String Operations) to make the Chg_Case function available to the outside world. The /Copy QCPYLESRC,STR_OPS_PR line that you see near the top of the code is providing my test program (TST_STROPS) with the function's prototype. Then I'm using all the different P_Case choices to produce different results.


You can play around with it a bit by downloading and running the source code for the Chg_Case function. As a bonus, the STR_OPS and TST_STROPS source members also include the code for the LEFT and RIGHT functions I explained a few TechTips ago.


These are just a few examples of Excel functions that you convert to RPG. You can argue that there's not much to gain from that. Well, at least you get to practice the whole process of creating a function! Many readers reacted to the use of %XLATE, saying that there are other ways to provide the same functionality. The purpose of these two TechTips was not promoting the use of %XLATE. It was simply using this BIF as a way to create functions in an easily understandable fashion. The SQL 101 TechTip series will present, in a few articles, SQL embedded functions that replace %XLATE with more reliable and functional SQL functions.


I'll continue to explore RPG's BIFs in the next TechTip. I'll be talking about data and time operations. Until then, share your thoughts, questions, and doubts with your fellow RPGers via the Comments section below or in the usual LinkedIn groups where these TechTips usually end up.