StringRegExp or ???

[funnydawg]

I continually receive the message:

Subscript used with non-Array variable.:

I do not know if this is related to StringRegExp or arrays, or both. Code is:

$asResult = StringRegExp($line, '((?i)[:alpha](1,25))', 1)

MsgBox(0, "Line Reads", $asResult[0])

The $line is a one line text string that starts with a bunch of 'junk' such as --99 48/56 and then gets to the part that I want which is alpha until the very end which always ends with 2-5 numbers.

Ideally I would like to capture from the start of the text - ignoring the junk - all the way to the end of the line.

unfortunately I continually get the error message mentioned above. Anyone know what I am doing wrong?

[PsaltyDS]

Just test for @error after StringRegExp() and don't try to use the returned array unless @error = 0:

$sLine = ""
$sPatt = '((?i)[:alpha](1,25))' ; Fix your pattern here
$asResult = StringRegExp($line, $sPatt, 1)
If @error Then
    MsgBox(16, "Error", "Error on RegExp; @error = " & @error & "; @extended = " & @extended)
Else
    MsgBox(0, "Line Reads", $asResult[0])
EndIf

Your pattern is still wrong, but you have the error handling now.

[zorphnog]

More than likely you have a bad expression. Anytime you are working with arrays you need to validate the data you receive for errors.

$asResult = StringRegExp($line, '((?i)[:alpha](1,25))', 1)
If @error Then
  MsgBox(0, "Line Reads", $asResult[0])
Else
  MsgBox(0, "Line Reads", "Error: No matches.")
EndIf

Edit: @PsaltyDS beat me to it.

Edited April 26, 2010 by zorphnog

[jchd]

Some samples of data to match would have been a must. Guesswork: "(?i)^[^a-z]*([a-z](?:\w|\s)*\d{2,5})$"

[GEOSoft]

"(?i).+([a-z]+\d{2,5}$"

[jchd]

I was about posting the same, but finally thought the "all text"part could well contain embedded digits, like below, so I added the extra guard.

-- 46 148906/246514896 Exception 17 occured at flight second 1645

[GEOSoft]

I was about posting the same, but finally thought the "all text"part could well contain embedded digits, like below, so I added the extra guard.

-- 46 148906/246514896 Exception 17 occured at flight second 1645

Yours will work just fine. If the "all text" was going to include digits then I would have used either \w+ or [a-z0-9].

The only reason I posted at all is, just the other day I was posting on my blog about the fact REGEX is great for solving complex problems but then we turn around and make the SRE too complex, or at least more complex than is usually required for the task at hand. I shouldn't make them that simple in the forums because we know that people seldom give us true examples of what they are starting with and what they expect returned.

[jchd]

True and many scripters with little experience see the need for precision and rigor as something which doesn't apply to their "simple" case. Unfortunately (or perhaps fortunately) their computers behave the same way as ours and deliver the same garbage as ours when directed by faulty logic or inconsistent rules. Today's PC simply do that faster.

[GEOSoft]

True and many scripters with little experience see the need for precision and rigor as something which doesn't apply to their "simple" case. Unfortunately (or perhaps fortunately) their computers behave the same way as ours and deliver the same garbage as ours when directed by faulty logic or inconsistent rules. Today's PC simply do that faster.

Isn't it amazing what people can class as simple when they are asking for help though? It can range from the truly simple to something that needs a super computer to do the calculations.

[jchd]

Fortunately we have a super language at our disposal, an incredibly wide-spectrum library, a huge cumulated culture and experience and above all ... big & brave hearts!

[GEOSoft]

Fortunately we have a super language at our disposal, an incredibly wide-spectrum library, a huge cumulated culture and experience and above all ... big & brave hearts!

You left out the "all mixed together with a pinch of stupidity."