Regex, TextReader...?

**Eric Gunnerson [MS]** · Jul 21 '05, 12:19 PM

Re: Regex, TextReader...?

Yes, you can do it in regex. The trick is to allow your pattern to match
more than one time. For example, if I had something like:

1234
34123
11313
113133
xxxxx

I could write something like:

(?<Numbers>^\d+ $)+xxxxx

Which means that I need to look at Match.Captures instead of Match.Groups,
IIRC.

Note that in most uses of this technique, what you really need to write is
something like:

((?<Numbers> match numbers) match stuff between numbers)+xxxxx

so that the match can continue. You may also need to play around with the
singleline and multiline options.

--
Eric Gunnerson

Visit the C# product team at http://www.csharp.net
Eric's blog is at http://weblogs.asp.net/ericgu/

This posting is provided "AS IS" with no warranties, and confers no rights.
"Masahiro Ito" <masa@pleasespa mgoaway.it> wrote in message
news:Xns947D7DE CA8DE7fujicomxy z@216.196.105.1 30...[color=blue]
> I have attached a block of text similar to the type that I am working
> with.
>
> I have been learning a lot about Regex - it is quite impressive. I can
> easily capture bits of info, but I keep having trouble with line breaks.
>
> I want to identify the start and end of blocks of text. Are there some
> tips someone can share?
>
> EG: in my text, I can grab a collection of everyones Phone number with:
> ^"M:"\t"(?<Phon eNumber>[^"])"
>
> But, what about if I wanted to grab many lines, until it matched a
> certain pattern. I use the ^ to say not the quote, but can I say not 14
> hyphens?
>
> The way I have split this type of data is inefficient. I match all the
> cases of:
> ^-{14}
> Then I use many math equations to split the file using the index of the
> matches. I am sure Regex must have some way to pattern match a complex
> not, to indicate the end of my match?
>
> Thank you.
>
>
>
>
> --------------
> "M:" "3242310532 "
> "Subscriber Name:" "MR Regex"
> "Additional line user name:" ""
> "Sublevel:" " "
> "Sublevel:" ""
> "Reference 1:" ""
> "Reference 2:" ""
>
> "CURRENT CHARGES"
> "Monthly Service Plan" $40.00
> "Additional Local Airtime" $0.00
> "Long Distance Charges" $0.00
> "Roaming Charges" $0.00
> "Network and Licensing Charges" $7.20
> "Total Taxes:" $7.09
> "Total Current Charges:" $47.20
>
> "MONTHLY SERVICE PLAN" 11-Oct-03 to 10-Nov-03
> "Service Plan Name" "Total"
> "Mike Dispatch 40 (11-Oct-03 to 10-Nov-03)" $40.00
> "Total Monthly Service Plan Charges" $40.00
>
> "ADDITIONAL LOCAL AIRTIME"
> "Service" "Total Mins. Used" "Free Mins. Used" "Included Mins.
> Used" "Chargeable Mins. Used" "Total"
> "Direct Connect Private (minutes)" 28:04 28:04 0:00 0:00 $0.00
> "Total Additional Local Airtime Charges" $0.00
>
> "LONG DISTANCE, ROAMING AND OTHER CALL CHARGES"
> "Service" "Incl. LD Minutes" "Chargeable LD Minutes" "Total"
> "Total Long Distance Charges" $0.00
>
> "ROAMING"
> "Service" "Roaming Minutes" "Roaming Charges" "Roaming LD Minutes"
> "Roaming LD Charges" "Roaming Surcharge" "Total"
> "Total Roaming Charges" $0.00
>
> "WIRELESS WEB - PREMIUM SERVICE"
> "Service" "Total Events" "Event Type" "Total"
> "Total Wireless Web Premium Services Charges" $0.00
>
> "PHONE - PREMIUM SERVICE"
> "Service" "Total Events" "Event Type" "Total"
> "Total Phone Premium Services Charges" $0.00
>
> "PAGER SERVICES"
> "Service" "Total Messages" "Included Messages" "Chargeable
> Messages" "Total"
> "Total Pager Charges" $0.00
>
> "VALUE-ADDED SERVICES" 11-Oct-03 to 10-Nov-03
> "Service" "Total"
> "Wireless Web - Surf Sampler (11-Oct-03 to 10-Nov-03)" $0.00
> "Total Value Added Service Charges" $0.00
>
> "OTHER CHARGES AND CREDIT"
> "Charge or Credit" "Total"
> "Total Other Charges and Credits" $0.00
>
> "NETWORK and LICENSING CHARGES"
> "Service" "Total"
> "911 Emergency Access Charge (11-Oct-03 to 10-Nov-03)" $0.25
> "System Licensing Charge (11-Oct-03 to 10-Nov-03)" $6.95
> "Total Network Licensing Charges" $7.20
>
> "TAXES"
> "" "Total"
> "Total Taxes" $7.09
>
> --------------
> "M:" "9042437121 "
> "Subscriber Name:" "Fred 1"
> "Additional line user name:" ""
> "Sublevel:" " "
> "Sublevel:" ""
> "Reference 1:" ""
> "Reference 2:" ""
>
> "CURRENT CHARGES"[/color]

**Masahiro Ito** · Jul 21 '05, 12:19 PM

Re: Regex, TextReader...?

Thank you Eric. I was doing a capture group (in my first example using
my sample text I used (?<PhoneNumber>[^"]*) to capture everything until
the next " in my phonenumber collection.

In this simple example, capturing the Field 1 and Field5 value, I cannot
reliably regex the 'everything between numbers'.

My attempt (doesn't work:
Field1:\s(<F1>[0-9]*)[^Field5:]*Field5:\s(?<F5 >[0-9.$]*)
^trouble^

Field1: 1234
Field2: 34123
Field3: 1313
Field4: 13133
Field5: $xxxx.00
Field6: 2342df
Field1: 2342
Field2: 33241
Field3: 2142
Field4: 543523
Field5: $342.00
Field6: 43254
Field1: 3415
Field2: 234235
Field3: 341
Field4: 13212533
Field5: $5234.00
Field6: 32415

Of course, I can run two separate captures, but...

You gave the example technique : ((?<Numbers> match numbers) match stuff
between numbers)+xxxxx

Does this +xxxxx match everything until the xxxxx is found? In my regex
apps (I use expresso and Regex Workshop as dotnet tools) there are no
matches.

Thanks,

Masa

"Eric Gunnerson [MS]" <ericgu@online. microsoft.com> wrote in
news:#Oygz5e5DH A.1804@TK2MSFTN GP12.phx.gbl:
[color=blue]
> Yes, you can do it in regex. The trick is to allow your pattern to
> match more than one time. For example, if I had something like:
>
> 1234
> 34123
> 11313
> 113133
> xxxxx
>
> I could write something like:
>
> (?<Numbers>^\d+ $)+xxxxx
>
> Which means that I need to look at Match.Captures instead of
> Match.Groups, IIRC.
>
> Note that in most uses of this technique, what you really need to
> write is something like:
>
> ((?<Numbers> match numbers) match stuff between numbers)+xxxxx
>
> so that the match can continue. You may also need to play around with
> the singleline and multiline options.
>[/color]

**Eric Gunnerson [MS]** · Jul 21 '05, 12:22 PM

Re: Regex, TextReader...?

I'm a little confused about what you're trying to do. Given the example text
below, what is the expect output that you want?

If I assume that you didn't mean to write xxxx.00 for the Field5 value
below, the following regex may do what you want:

new Regex(@"
(
(?<S2>.*?)
Field1:\s(?<F1>[0-9]*)
(?<S1>.+?)
Field5:\s(?<F5>[0-9.\$]+)
)+",
RegexOption.Ign orePatternWhite space);

All the F1 values will be in one capture, all the F5 values in the other
capture. I named the S1 and S2 captures so you could see what they're
matching.

I'd suggest using my Regex Workbench at
http://www.gotdotnet.com/Community/U...1-4EE2729D7322 -
it makes playing around with Regex much easier.

--
Eric Gunnerson

Visit the C# product team at http://www.csharp.net
Eric's blog is at http://weblogs.asp.net/ericgu/

This posting is provided "AS IS" with no warranties, and confers no rights.
"Masahiro Ito" <masa@pleasespa mgoaway.it> wrote in message
news:Xns947F55B 615E9fujicomxyz @207.46.248.16. ..[color=blue]
> Thank you Eric. I was doing a capture group (in my first example using
> my sample text I used (?<PhoneNumber>[^"]*) to capture everything until
> the next " in my phonenumber collection.
>
> In this simple example, capturing the Field 1 and Field5 value, I cannot
> reliably regex the 'everything between numbers'.
>
> My attempt (doesn't work:
> Field1:\s(<F1>[0-9]*)[^Field5:]*Field5:\s(?<F5 >[0-9.$]*)
> ^trouble^
>
> Field1: 1234
> Field2: 34123
> Field3: 1313
> Field4: 13133
> Field5: $xxxx.00
> Field6: 2342df
> Field1: 2342
> Field2: 33241
> Field3: 2142
> Field4: 543523
> Field5: $342.00
> Field6: 43254
> Field1: 3415
> Field2: 234235
> Field3: 341
> Field4: 13212533
> Field5: $5234.00
> Field6: 32415
>
> Of course, I can run two separate captures, but...
>
> You gave the example technique : ((?<Numbers> match numbers) match stuff
> between numbers)+xxxxx
>
> Does this +xxxxx match everything until the xxxxx is found? In my regex
> apps (I use expresso and Regex Workshop as dotnet tools) there are no
> matches.
>
> Thanks,
>
> Masa
>
>
>
> "Eric Gunnerson [MS]" <ericgu@online. microsoft.com> wrote in
> news:#Oygz5e5DH A.1804@TK2MSFTN GP12.phx.gbl:
>[color=green]
> > Yes, you can do it in regex. The trick is to allow your pattern to
> > match more than one time. For example, if I had something like:
> >
> > 1234
> > 34123
> > 11313
> > 113133
> > xxxxx
> >
> > I could write something like:
> >
> > (?<Numbers>^\d+ $)+xxxxx
> >
> > Which means that I need to look at Match.Captures instead of
> > Match.Groups, IIRC.
> >
> > Note that in most uses of this technique, what you really need to
> > write is something like:
> >
> > ((?<Numbers> match numbers) match stuff between numbers)+xxxxx
> >
> > so that the match can continue. You may also need to play around with
> > the singleline and multiline options.
> >[/color]
>[/color]

**Masahiro Ito** · Jul 21 '05, 12:26 PM

Re: Regex, TextReader...?

"Eric Gunnerson [MS]" <ericgu@online. microsoft.com> wrote in
news:OQ1JlL55DH A.2496@TK2MSFTN GP09.phx.gbl:
[color=blue]
> I'm a little confused about what you're trying to do. Given the
> example text below, what is the expect output that you want?
>
> If I assume that you didn't mean to write xxxx.00 for the Field5 value
> below, the following regex may do what you want:
>
> new Regex(@"
> (
> (?<S2>.*?)
> Field1:\s(?<F1>[0-9]*)
> (?<S1>.+?)
> Field5:\s(?<F5>[0-9.\$]+)
> )+",
> RegexOption.Ign orePatternWhite space);
>
> All the F1 values will be in one capture, all the F5 values in the
> other capture. I named the S1 and S2 captures so you could see what
> they're matching.
>
> I'd suggest using my Regex Workbench at
> http://www.gotdotnet.com/Community/U...px?SampleGuid=
> C712F2DF-B026-4D58-8961-4EE2729D7322 - it makes playing around with
> Regex much easier.[/color]

Thanks Eric. Actually, I was using your Regex Workbench already - it is
great! Thank you for sharing it.

Something is not clicking with me and these regex expressions. Even when I
paste your regex, I don't believe I am getting the responses you intended.
In the sample I posted, I am trying to capture the field 1 and field 5
values. I can capture them separately, but can't seem to grasp the 'skip
everything until a specific pattern is matched'.

I am trying to break down your sample piece by piece. Does the @ at the
start do something?

Also, using Regex Workbench, using your sample in your first reply, I am
not getting any matches.
String:
1234
34123
11313
113133
xxxxx

Regex:
(?<Numbers>^\d+ $)+xxxxx

I have tried every permutation I can think of with Multi/single line, etc..
I feel like I am going crazy.

Thank you.

Masa

Regex, TextReader...?

Regex, TextReader...?

Comment

Comment

Comment

Comment