Scan String Array

[Tom Cone Jr]

I'm having trouble with ARRAY SCAN.

I've created a string array, and I'm trying to find the array element that contains the character string "Date:"

This seems like such a basic thing, I'm embarrassed to post, but I've studied the manual and searched the message board... so here goes.

I'm using vers 9.01.

Here's my code:

#COMPILE EXE
#DIM ALL

FUNCTION PBMAIN () AS LONG

    DIM infile AS STRING

    DIM reg_date AS STRING

    DIM source_txt AS STRING

    DIM line_count AS LONG
    DIM arr_index&

    source_txt = ""
    source_txt = INPUTBOX$("Enter fully justified source TXT filename","Source file?")
    IF source_txt = "" THEN
        EXIT FUNCTION
    END IF

    OPEN source_txt FOR INPUT AS #1  'i.e. open message previously saved as text
    IF ERR THEN
        MSGBOX "Could not open source file",,"Unable to open file"
        EXIT FUNCTION
    END IF

    'get line count
    FILESCAN #1, RECORDS TO line_count

    'now read the lines of the message into an array
    DIM msgline(1 TO line_count) AS STRING
    LINE INPUT #1, msgline$() TO line_count  'populate the array
    CLOSE #1  'close source message text file
 
    'reg_date
    ARRAY SCAN msgline(), ="Date:", TO arr_index&  'find line containing the string date:
    MSGBOX "array index: " + STR$(arr_index)  'returns 0
    
    EXIT FUNCTION


END FUNCTION

The array is populated by reading a simple text file. Here's a sample:

MessageEarthLink.net | My Start Page | myVoice | My Account | Support |

            Message
            ‹ Previous | Next › | « Back to INBOX 
               Forward... inline text as attachment     Move to... Trash  More 
              Actions... View Headers View Source --------------- Mark Unread 
              Flag Message 
            From: 
            To: 
            Subject: 
            Date: Apr 23, 2009 6:10 PM
            ‹ Previous | Next › | « Back to INBOX 


© 2008 EarthLink, Inc. All Rights Reserved.
Members and visitors to the EarthLink Web site agree to abide by our Policies 
and Agreements
EarthLink Privacy Policy 
Web Mail version 6.0.32

When my code is run the arr_index that is returned is zero.

What am I doing wrong?

Thanks.

-- tom

[Rodney Hicks]

Perhaps a typo:

 MSGBOX "array index: " + STR$(arr_index)

should read

 MSGBOX "array index: " + STR$(arr_index&)

[Tom Cone Jr]

Rod, thanks, but no joy. The Array Scan statement is returning zero. It can't find the search string, even after I correct the typo you mention.

-- tom

[Michael Mattias]

ARRAY SCAN msgline(), ="Date:", TO arr_index&  'find line containing the string

This statement does not find an element containing the string, it finds an element which equals the string.

MCM

[Tom Cone Jr]

Thanks, Michael.

The language in the help file that confused me says that this statement scans all or part of an array for a given value. I misconstrued what this means.

Since I'm searching for a substring within a string array I suppose I'll need to loop through the elements of the array using INSTR() to see if the search string is found within any of them. This will suffice for my needs, but I'm curious. Might there be a better way?

Thanks again.

-- tom

[John Gleason]

I think that's what you need to do, the trusty loop function:

FUNCTION ArrayContains(arr() AS STRING, findString AS STRING) AS LONG
   LOCAL ii AS LONG
   
   FOR ii = LBOUND(arr()) TO UBOUND(arr())
      IF INSTR(arr(ii), findString) THEN
         FUNCTION = ii
         EXIT FUNCTION
      END IF
   NEXT
END FUNCTION

then your code might check for the return:

    findLine = ArrayContains(msgline(), "Date:")
    IF findLine THEN ? "Found at line" & STR$(findLine)

[John Gleason]

Might there be a better way?

I did a quick TIX check on the loop function's speed and it's about 15 tix per array byte. A 2GHz machine will check ~ 133 MB/sec--quite speedy.

[Tom Cone Jr]

Thanks!

[Rodney Hicks]

Looking at the supplied code and data and if you reread what Micheal posted you will see that the following, at least as I read it.

LINE INPUT #1, msgline$() TO line_count

will put into some element a line thus:

Date: Apr 23, 2009 6:10 PM     'this is what the array scan finds, not DATE:

which means there will be no line in the msgline$() that matches the string in the ARRAY SCAN statement:

ARRAY SCAN msgline(), ="Date:", TO arr_index&

which takes you to this line from the help on ARRAY SCAN:

If none of the scanned elements satisfy expression, zero will be stored in lvar&

which is exactly the result you're getting.

You might get what you want faster by doing:

    FOR xx=1 TO line_count
      IF LEFT$(msgline$(xx),5)="Date:" THEN arr_index& = xx
    NEXT xx

Mr Gleason's code does the same thing in a different manner, but now you should know why you were getting the result you were getting.

[Tom Cone Jr]

Thanks, Rod. I appreciate the further explanation and example. The documentation is plain enough. I'm surprised I didn't understand it sooner. I suppose I was going too fast, and having hit the problem failed to slow down and actually read the material carefully. Jumped to the conclusion that my implementation of the function was flawed. Lesson learned (at least for a while). Thanks again.

-- tom

[Michael Mattias]

???

ARRAY SCAN stringarray([start] [FOR COUNT] , _ 
        =|>[=]|<[=]]| LIKE {stringvar|stringliteral} [COLLATE UCASE] TO IHit

?????

MCM

[Dave Biggs]

When scanning a string array, all characters of each element of the array are normally considered when performing comparisons. To limit the comparison to a specific subset of characters, use FROM to specify the start position, and TO to specify the end position that ARRAY SCAN will consider within each array element...

If the character positions always hold true to your sample data you could search with..

ARRAY SCAN msgline(), From 13 to 17, Collate UCASE, ="DATE:", TO arr_index&  'find line containing the string date:

[John Petty]

You could also concatanate the entire array into a single string with JOIN$ then just use one INSTR. If found then you would only have to add up the string lengths of the array members + 1 for the delimeter until you reach the found number so giving you the array member

[Michael Mattias]

found then you would only have to add up the string lengths of the array members + 1 for the delimeter until you reach the found number so giving you the array member

You wouldn't even have to get all those individual LEN() results if you've created a master string using JOIN$......

Once you get a hit with INSTR, just count the number of delimiters found in that JOINed string up to the point of the hit....

LOCAL pb AS BYTE PTR , bdelim AS BYTE 

   S = (JOINED string using delimiter) 
   bDelim = ASC(delimiter) 
   z = INSTR (target, s) 
   pb  = STRPTR(s) 
   IF Z then 
       ichar  = 0 
       DO UNTIL ichar = Z  
            IF @pb[ichar] = bDelim then
               INCR  element_no 
            END IF 
            INCR  iChar 
      LOOP 
 END IF

(Something like that anyway).
... because the number of delimiters found in the JOINed string *is* an "element counter."

MCM

[John Petty]

Good point, I have no idea which would be faster, adding the the longs that are stored in every string desciptor or as you have shown a good use of pointers