regex again...

[Aceguy]

hi guys... stuck on this regex....

i want to validate this line of text with regex....

Please be advised that the above invoice has been dispatched to the current address

so i want to validate it by has|have and the word dispatched and it MUST contain BOTH instances.... has dispatched | have dispatched...

i have tried

(has|have)dispatched

but it only validates dispatched, and ignnores has|have....

so basically the dialog must contain (has or have) and must also contain (dispatched)

Edited February 17, 2010 by Aceguy

[notsure]

Im not sure what you're trying to do. Please rephrase the problem...

You want "Dispatched" how?

Something like this:

Please be advised that the above invoice has been has dispatched|have dispatched to the current address

or what? Please be more clear about what you're trying to do.

[Aceguy]

the dialog must contain (has or have) and must also contain (dispatched)

[notsure]

Its better to validate the line with stringinstr then i think. I dont know if i get it still,... but you want to be sure HAS OR HAVE AND DISPATCHED are in the string?

[Aceguy]

must be regex...

but you want to be sure HAS OR HAVE AND DISPATCHED are in the string?

that is correct

Edited February 17, 2010 by Aceguy

[notsure]

#include <Constants.au3>

local $string="Please be advised that the above invoice have been dispatched to the current address"
if StringRegExp($string, "(?i:.*has|have.*dispatched.*)") = 1 Then
msgbox(1,"","got it.")

EndIf

[99ojo]

Hi,

notsure points the way, but it isn't the goal you want:

$string0 = "Match:Please be advised that the above invoice has been dispatched to the current address"
$string1 = "Match:Please be advised that the above invoice have been dispatched to the current address"
$string2 = "No Match:Please be advised that the above invoice have been dispatchd to the current address"
$string3 = "No Match:Please be advised that the above invoice has been dispatchd to the current address"
$string4 = "No Match:Please be advised that the above invoice hav been dispatched to the current address"
For $i = 0 To 4
    $match = StringRegExp (Eval ("string" & $i), "(?i:.*has|have)(?i:.*dispatched.*)")
    Consolewrite ($match & " " & Eval ("string" & $i) & @CRLF)
Next

;-))

Stefan

Edited February 17, 2010 by 99ojo

[notsure]

Then i still don't get the goal.. He wants to be sure have/has is in the string AND dispatched is in the string. Well, in my code they are and then enters the If/Then statement.

I'm confused ;(

[99ojo]

Then i still don't get the goal.. He wants to be sure have/has is in the string AND dispatched is in the string. Well, in my code they are and then enters the If/Then statement.

I'm confused ;(

Hi,

Your string matches your pattern, thats right, but fits only some combinations of (have or has) And (dispatched). See my test strings.

Just take my code and put your pattern in.

You will see, that $match is 1 on $string3, which isn't the goal.

;-))

Stefan

Edited February 17, 2010 by 99ojo

[notsure]

Hi,

Your string matches your pattern, thats right, but fits only some combinations of (have or has) And (dispatched). See my test strings.

Just take my code and put your pattern in.

You will see, that $match is 1 on $string3, which isn't the goal.

;-))

Stefan

Ah ok... thanks for the explanation, you're right.

[Aceguy]

thanks guys for your input, both have been greatfullly appreciated....

thanks again.... i never have got the hang of regex.

Edited February 17, 2010 by Aceguy

[dani]

@Aceguy

It's really not that difficult. You were very close but have to realize the input string ("Please be advised that the above invoice has been dispatched to the current address") has to match the *entire* regexp. In your case, that only was '(has|have)dispatched'. An input sentence like "hasdispatched" or "havedispatched" would have matched that. However, wildcards are needed to match the *other* characters. That's why you need .* to pad it here and there (. matches any character, * = 0 or more times). The ?i makes it case insensitive.

[99ojo]

@Aceguy

It's really not that difficult. You were very close but have to realize the input string ("Please be advised that the above invoice has been dispatched to the current address") has to match the *entire* regexp. In your case, that only was '(has|have)dispatched'. An input sentence like "hasdispatched" or "havedispatched" would have matched that. However, wildcards are needed to match the *other* characters. That's why you need .* to pad it here and there (. matches any character, * = 0 or more times). The ?i makes it case insensitive.

Hi,

thanks for further excurse to regexp.

I started only with the suggestion from @notsure and did the try and error method and find as result the solution.

But i couldn't explain, why @notsures code doesn't match the goal.

;-))

Stefan

Edited February 18, 2010 by 99ojo

[jchd]

But i couldn't explain, why @notsures code doesn't match the goal.

Let's look at it. The pattern was "(has|have)dispatched". To go thru regexps, put them in plain language, e.g. english (I can't write german for such things).

The pattern will match an input containing one of the sequences "has" or "have" immediately followed by the sequence "dispatched".

You now "see" that you need to get rid of the "immediately" part. You can do so by inserting the pattern ".+" meaning that "has" or "have" have to be separated by at least one character for the whole pattern to match. You then end up with "(has|have).+dispatched" which works.

Even if that last patterns works, I find it a bit imprecise. At least, the words we're after should be treated like words, this way substrings in "shave" or "phasing out" won't produce false positives if ever the phrasing changes: "(?i)\bha(s|ve)\b.*\bdispatched\b"

[99ojo]

Let's look at it. The pattern was "(has|have)dispatched". To go thru regexps, put them in plain language, e.g. english (I can't write german for such things).

The pattern will match an input containing one of the sequences "has" or "have" immediately followed by the sequence "dispatched".

You now "see" that you need to get rid of the "immediately" part. You can do so by inserting the pattern ".+" meaning that "has" or "have" have to be separated by at least one character for the whole pattern to match. You then end up with "(has|have).+dispatched" which works.

Even if that last patterns works, I find it a bit imprecise. At least, the words we're after should be treated like words, this way substrings in "shave" or "phasing out" won't produce false positives if ever the phrasing changes: "(?i)\bha(s|ve)\b.*\bdispatched\b"

Hi,

good explanation. Makes it more and more clear.

Thanks a lot.

;-))

Stefan