Software Du Jour

home *** CD-ROM | disk | FTP | other *** search

/ Software Du Jour / SoftwareDuJour.iso / BUSINESS / DBASE / CLIPUDF.ARC / FIRSTCAP.PRG < prev next >

Wrap

Text File | 1986-02-04 | 14.9 KB | 419 lines

EDITORS NOTE: NPN member Denny Dias has supplied us with a handy Clipper UDF (User Defined Function). As the our first publication of an outside article, I would like to encourage all potential NPN authors to check out Denny's formating and composition. This is the type of 'copy' that editors dream of. Thanks Denny! Because of the length of this article, and because I am presuming that most people will not be reading it online, there will not be any page breaks from this page on. If you do intend on reading the article online, use the CTRL S and CTRL Q keys to stop and start the text. If you wish to leave the article type CTRL P. So here comes FIRSTCAP. Fire it up...its fast...you'll love it! -ROGER FIRSTCAP . . . THE FUNCTION --------------------------- by Dennis L. Dias (NAN449) Sometimes it is necessary to capitalize the first character of each word in a character string. This is true for titles, headings, and proper names. There is a procedure listed on the Ashton-Tate network that has as similar purpose; however that routine has shortcomings that inspired me to write my own. For one thing, the Ashton-Tate program is a procedure not a function; therefore it can only return a result by altering its input. Consequently, it can only recieve a memory variable as input. Another problem is the use of the AT() function to find the next space between words. That algorithm will return "Ashton-tate" (the 't' is left uncapitalized) and "C&h Sugar". Admittedly, the routines presented here are not perfect. There is no test for context or grammer. However, most of the problems have been resolved and the function may be used in many ways. For example: LIST FIRSTCAP(TRIM(last) + ", " + TRIM(first) + " " + TRIM(middle)) The main problem with the first version of FIRSTCAP() is speed. The routine is much too slow..so slow that I would hesitate to use it for anything more than a heading. Repeated calls are out of the question. It is possible to gain a little speed by capitalizing the first word before entering the main loop and eliminating the ( LEN() > 0 ) test in the first CASE statement. However, that is not enough to permit the repeated calls of a LIST command. This led me to the second version. In the second version I make use of assembly language to access the full speed of the CPU. This version is so fast that it is completely invisible during program execution. It also demonstrates one method of using assembly language within a user defined function. Here, then, are both versions of FIRSTCAP() along with the subfunction for version 1 and assembler routine necessary to utilize version 2. ********************************************* * V E R S I O N 1 * ********************************************* * * Function.....: Firstcap() . . . Version 1 * Author.......: Dennis L. Dias * Source ID....: NAN449 * Date.........: 12/14/85 * * Syntax: FIRSTCAP( <expC> ) * Return: Character string with first letter of each word * capitalized and the rest lower case. The special * words "and", "but", "for", "the", "by", "an", "of", * "on", "in", "if", "to", "or", "at" and "a" will remain * lower case unless: (A) The word is the first word in * the string with no leading word separators; (B) The * word is followed by a non-space character or (C) The * word is the last word in the string. * * Notes: This first version is written entirely in dBASE. It * calls upon a second function (ISCHAR()..also written * in dBASE) to determine which characters are "in-word" * characters and which are not. The execution speed is * slow..unacceptable for repeated calls. * FUNCTION FIRSTCAP * PARAMETERS fl_strg PRIVATE fl_begn,fl_len,fl_pos,fl_outstr,fl_part,fl_true * fl_begn = 0 fl_len = 0 fl_pos = 1 IF LEN(fl_strg) = 1 fl_outstr = UPPER(fl_strg) ELSE fl_outstr = "" ENDIF fl_part = "" fl_true = LEN(fl_strg) > 1 DO WHILE fl_true fl_begn = fl_pos * * Scan past in-word characters * DO WHILE ISCHAR(SUBSTR(fl_strg,fl_pos,1)) .AND. fl_pos < LEN(fl_strg) fl_pos = fl_pos + 1 ENDDO * * Scan past word separators * DO WHILE .NOT. ISCHAR(SUBSTR(fl_strg,fl_pos,1)) .AND.; fl_pos < LEN(fl_strg) fl_pos = fl_pos + 1 ENDDO * * Determine if this is the last word * IF fl_pos = LEN(fl_strg) .AND. (ISCHAR(SUBSTR(fl_strg,fl_pos - 1,1)); .OR. fl_begn = fl_pos) fl_len = fl_pos - fl_begn + 1 * * Set false to exit main loop * fl_true = .F. ELSE fl_len = fl_pos - fl_begn ENDIF * * Isolate one word and convert it to lower case * fl_part = LOWER(SUBSTR(fl_strg,fl_begn,fl_len)) * * Test for special word or a one character word at the end of * the line. The use of (" " + TRIM() + " ") protects against * an unwanted substring match such as ("he " $ "the "). * DO CASE CASE (" " + TRIM(fl_part) + " ") $ " and the for but to or by an" +; " in of on at if a " .AND. LEN(fl_outstr) > 0 .AND.; fl_pos < LEN(fl_strg) * * Special word found in mid-string position where it should * not be capitalized..concatenate the word as is * fl_outstr = fl_outstr + fl_part CASE LEN(fl_part) = 1 * * LEN() = 1 only occurs when a one character word appears * at the end of the string..capitalize and concatenate * fl_outstr = fl_outstr + UPPER(fl_part) OTHERWISE * * Capitalize the first character and concatenate * the word to the output string * fl_outstr = fl_outstr + UPPER(SUBSTR(fl_part,1,1)) +; SUBSTR(fl_part,2) ENDCASE ENDDO RETURN(fl_outstr) * ********************************** ********************************** * * Function.....: Ischar() * Author.......: Dennis L. Dias * Source ID....: NAN449 * Date.........: 12/14/85 * * Syntax: ISCHAR( <expC> ) * Return: Logical true if the first character in <expC> is alpha * or STR() or an apostrophe (') * FUNCTION ISCHAR * PARAMETERS fl_string * RETURN(UPPER(SUBSTR(fl_string,1,1)) $; "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ'") * ********************************************* * V E R S I O N 2 * ********************************************* * * Function.....: Firstcap() . . . Version 2 * Author.......: Dennis L. Dias * Source ID....: NAN449 * Date.........: 12/14/85 * * Syntax: FIRSTCAP( <expC> ) * Same syntax and output as version 1 except that the first word * will always be capitalized..even if a special word is preceded * by one or more leading word separators such as "...A Fine * Madness...". * * Notes: This version calls an assembly language routine to do * most of the work. The character string passed to the * assembler module must first be converted to lower case. * Execution speed is very fast..completely invisible. * FUNCTION FIRSTCAP * PARAMETERS fl_strg PRIVATE fl_outstr * fl_outstr = LOWER(fl_strg) CALL fcap with fl_outstr RETURN(fl_outstr) * *************************** ;-------------------------------: ; Program File.: FCAP.ASM : ; Author.......: Dennis L. Dias : ; Source ID....: NAN449 : ; Date.........: 12/14/85 : ;-------------------------------: ; ;-----------------------------------------------------------------: ;This assembly language routine is to be called by Clipper to : ; convert the first letter of each word of the input string to : ; caps. The special words listed as "special" below will remain : ; lower case unless they appear as the first or last word or if : ; they are followed by a non- space character. The input string : ; must be passed as lower case. The apostrophe (') is : ; considered a character. : ;-----------------------------------------------------------------: ; public fcap ;Public procedure ; _prog segment byte ;Clipper segment assume cs:_prog,ds:_prog,es:nothing,ss:nothing ; ;----------------: ;Local data area : ;----------------: ; ;Special words are grouped according to length. Each group is ; preceeded by the length plus one space, and the number of words ; of that length. This greatly reduces the number of comparisons ; required to search the entire list. These numbers MUST be ; changed when adding or removing wods from the list. ; special db 4,4,"and the for but " db 3,9,"to or by an in of on if at " db 2,1,"a ",0 ;Special words are terminated with null ; ;List of characters to be considered "in-word" characters ; char db "0123456789abcdefghijklmnopqrstuvwxyz'" ; ;------------------------------------------------------: ;Primary routine..far procedure as required by Clipper : ;------------------------------------------------------: ; fcap proc far ; push bp ;Required by Clipper mov bp,sp ;To address variable at SP + 6 push ds ;Must save seg regs push es cld ;Forward direction flag ; lds si,dword ptr[bp+6] ;Point to input string with DS:SI push cs pop es ;Address local data via ES register ; lea di,char ;ES:DI points to in-word characters mov cx,37 ;Length of in-word characters ; ;Capitalize the first word no matter what it is ; fc1: lodsb ;Fetch char or al,al ;Check for null terminator jz fc_ret ;Quit if null..no characters at all ; call fl_scan ;Test for in-word character jnz fc1 ;Scan off word separators mov bx,si dec bx ;DS:BX points to first character call fl_cap ;Cap first letter ; ;Return here for each new word ; fc2: mov bx,si dec bx ;DS:BX points to first character ; fc3: lodsb ;Fetch character or al,al ;Check for null terminator jz fc5 ;Cap last word ; call fl_scan ;Test for in-word character jz fc3 ;Scan past in-word characters ; fc4: mov dx,si sub dx,bx ;DX has size of word plus 1 cmp dl,4 ;Special words are 3 chars ja fc5 ; plus 1 space max ; call fl_com ;Test for special word jc fc6 ;Found..don't cap ; fc5: call fl_cap ;Cap first letter ; fc6: dec si ;Point to end of prev word + 1 lea di,char ;Point to in-word characters mov cx,37 ;Length of in-word characters ; fc7: lodsb ;Fetch char or al,al ;Check for null terminator jz fc_ret ;Quit if null ; call fl_scan jnz fc7 ;Scan off word separators jmp fc2 ;Go process next word ; ;Done ; fc_ret: pop es ;Restore registers for Clipper pop ds pop bp ret ;Far return to Clipper ; fcap endp ; ;------------------: ;Local subroutines : ;------------------: ; ;The following procedure tests the character in AL as being in-word or ; not. The zero flag returns set if the character is in-word. ; fl_scan proc near ; push di ;Save pointer to in-word characters push cx ;Save length of in-word characters repne scasb ;Look for match pop cx ;Restore saved values and return pop di ; result in Zero flag ret ;Near return ; fl_scan endp ; ;Convert lower case to upper case..DS:BX points to the character ; fl_cap proc near ; cmp byte ptr[bx],"a" jb fl_cap_ret ;Less than "a" not lower cmp byte ptr[bx],"z" ja fl_cap_ret ;Above "z" not lower sub byte ptr[bx],20h ;Convert to upper case ; fl_cap_ret: ret ;Near return ; fl_cap endp ; ;Look for special words..return carry flag set if found..on entry ; DS:BX points to first character of the word to test. Begin by ; exchanging the contents of ES with DS to facilitate the use of ; the string instructions. AX, CX, DX, and DI destroyed. ; fl_com proc near ; push si ;Save current pointer push ds pop es ;Clipper segment to ES push cs pop ds ;Local segment to DS mov di,bx ;ES:DI points to current word lea si,special ;DS:SI points to local list of words ; fl_c1: lodsb ;Fetch word size or al,al ;Check for null terminator jz fl_clc ;Exit routine..special word not found ; mov cl,al ;Size of next group of words to count reg xor ch,ch ;Zero high byte lodsb ;Fetch word count mov dh,al ;To DH for outer loop mov ax,cx ;Save count in AX ; fl_c2: push di ;Save pointer to test word mov cx,ax ;Word size to counter repe cmpsb ;Compare bytes je fl_stc ;Jump if special word found add si,cx ;Point to next word or next group pop di ;Recover pointer to test word dec dh ;Reduce outer loop counter jnz fl_c2 ;Anoter word..same length jmp fl_c1 ;Go for next group of words ; fl_clc: clc ;Clear carry..special word not found ; fl_com_ret: pop si ;Restore registers and return push es pop ds ;Clipper segment to DS push cs pop es ;Local segment to ES ret ;Near return ; ;Special word found..return result in carry flag ; fl_stc: pop di ;Clear stack stc ;Set carry flag jmp fl_com_ret ;Restore and return ; fl_com endp _prog ends end