[Solved] Return strings that do NOT match regexp pattern

[kylomas]

Good Afternoon,

Given the following string :

03/03/2010
1/12/2010
ee/ee/ee
10/3/1999
22/22/2222
11/11/11

\d\d/\d\d/\d\d\d\d

will give me the two matches that I expect.

Is there a way to return all lines that do NOT match the pattern?

I've tried playng around with groups and the caret char to no avail, obviously.

Thanks,

kylomas

Edited October 12, 2012 by kylomas

[jdelaney]

best i can do:

$string = "03/03/2010" & @CRLF & _
"1/12/2010" & @CRLF & _
"ee/ee/ee" & @CRLF & _
"10/3/1999" & @CRLF & _
"22/22/2222" & @CRLF & _
"11/11/11"
$adata = StringRegExp ( @CRLF & $string & @CRLF, "sd/d+/d+|d+/d/d+|sd+/d+/d{1,3}s|s[^ds]+.+/[^ds]+.+/[^ds]+.+s", 3 )
_ArrayDisplay ( $adata )

eaiser with stringregexp replace:

$string = "03/03/2010" & @CRLF & _
"1/12/2010" & @CRLF & _
"ee/ee/ee" & @CRLF & _
"10/3/1999" & @CRLF & _
"22/22/2222" & @CRLF & _
"11/11/11"
$string2 = StringRegExpReplace ( $string, "d{2,2}/d{2,2}/d{4,4}", "" )
$adata = StringSplit ( $string2, @CRLF, 2 )
For $i = UBound ( $adata ) - 1 To 0 Step -1
 If StringLen ( $adata[$i] ) = 0 Then _ArrayDelete ( $adata, $i )
Next
ConsoleWrite ( $string2)
_ArrayDisplay ( $adata )

Edited October 12, 2012 by jdelaney

[kylomas]

jdelaney,

Thanks for the prompt reply!

If I understand you regexp correctly you are looking for all patterns that do not match the pattern that I want. This works for this example, however, I am auditing data prior to inclusion to a database and may not be able to id all possible BAD combinations. That is where my idea to return the NOT of matches came from.

I want to review the data that does not match what I expect to make sure that I am anticipating all cases.

kylomas

[kylomas]

jdelaney,

Sorry, missed your edit. I think that will work perfectly.

kylomas

edit: although I was hoping not to have to do this logically with arrays.

*** WARNING ***

The Surgeon General has determined that regular expression may be hazardous to your health.

Edited October 12, 2012 by kylomas

[jdelaney]

I just added the arrays to get rid of those remnant carriage returns...couldn't get rid of them with the regexp.

[jchd]

If you need to sanitize individual data only, then use a simple StringRegExp with the required pattern and option 0. This should give you the green/red light you want.

[jchd]

@jdelaney,

including a final s? in the pattern should work or am I missing something obvious?

[kylomas]

@jdelaney - thanks, works perfectly

@jchd - thanks, I am doing some of that already. The case that I'm working on now is weeding out the myriad of ways that a date can be entered.

I just started on this and it seemed that returning the NOT of an expression would be handy in many cases.

kylomas

[jdelaney]

@jchd, i would expect that too

even took into consideration that it was questionable s? (in case last record), no amount of manipulation could get rid of them, so i compromised on the array

Edited October 12, 2012 by jdelaney

[kylomas]

Jdelaney,

Your array method gets rid of the blank entries which makes the data easier to work with.

kylomas

[jdelaney]

duh, crlf is 2 characters, not 1...here is the way without the array:

$string = "03/03/2010" & @CRLF & _
"1/12/2010" & @CRLF & _
"ee/ee/ee" & @CRLF & _
"10/3/1999" & @CRLF & _
"22/22/2222" & @CRLF & _
"11/11/11"
$string2 = StringRegExpReplace ( $string, "d{2,2}/d{2,2}/d{4,4}[s]*", "" )
$adata = StringSplit ( $string2, @CRLF, 2 )

ConsoleWrite ( $string2)

[kylomas]

jdelaney,

Thanks, perfect and almost instantaneos tested against a 50KB file. Inj retrospect I do need to see the blank entries as they indicate missing data.

kylomas

edit: brain fart from exposure to regexp

Edited October 12, 2012 by kylomas