topic Re: Regular Expressions Board in BreakPoint

Regular Expressions Board

aeastet — Tue, 20 Jul 2010 12:15:23 GMT

We have an excel board. Should we have a repository for regular expressions too. This might be just and useful as the Excel board.

A Regular Expressions to check for prime numbers

jcarmody — Tue, 20 Jul 2010 23:33:20 GMT

I'll begin with this one that's more of a curiosity than a useful snippet.

I saw it here (with an explanation) and had to test it in LabVIEW. Someone is very clever.

What you can get out of this is an exercise in familiarity with the elements of Regular Expressions.

The pipe causes the regex to match either the expression on the left OR the one on the right
the carrot looks for a match at the beginning of the input string
the dollar sign looks for a match at the end of the input string
the question mark makes the preceding item optional (zero or one of whatever is before it)
beginning a regex with a carrot and ending it with a dollar sign only returns an exact match (in this case, either zero or one "1"s)
and then, the right side of the pipe...

Regular Expressions humor

jcarmody — Wed, 21 Jul 2010 10:35:26 GMT

Why doesn't Johnny use Regular Expressions?

His mother doesn't let him play with matches. 😄

Re: A Regular Expressions to check for prime numbers

Intaris — Wed, 21 Jul 2010 15:13:39 GMT

Took me a while to find the carrott.

I think you mean "caret" (pronounced same as carrott).....

Shane

Re: A Regular Expressions to check for prime numbers

jcarmody — Wed, 21 Jul 2010 16:53:26 GMT

EDIT: retracted.

not to be confused ...

altenbach — Wed, 21 Jul 2010 18:08:26 GMT

Not to be confused with irregular expressions. 😄

Re: not to be confused ...

tbob — Wed, 21 Jul 2010 18:24:43 GMT

@altenbach wrote:

Not to be confused with irregular expressions. 😄

You have to be an irregular guy to be able to understand irregular expressions.

an example of how Match Regular Expression beats Match Pattern

jcarmody — Fri, 23 Jul 2010 10:09:47 GMT

This example demonstrates the oft-used expression: ".*" (dot star). You remember (right?) that a dot matches "any single character except line break characters \r and \n" and that the star "[r]epeats the previous item zero or more times." The parentheses save whatever is matched by the expression between them (you can use this later in your expression as a "backreference"). In this example I'm looking for the word "and", but I don't want to have to strip off the "start", "stop" or the whitespace later. I only use the Whole Match output to illustrate the function; the string I'm interested in is returned in the first submatch. (Note: MRE is an expandable node. You need to drag the bottom handle down to access the submatch terminals.)

This method should replace the following snippet:

Please don't ever do this again. 😛

one way to use a backreference

jcarmody — Fri, 23 Jul 2010 22:18:56 GMT

Submatches are useful in extracting a part of your match but they also can also be reused in the regexp. An example is to match beginning and ending tags when you don't know what tag you're looking for.

The brackets are not special characters so they'll result in a literal match. The expression "(\w*)" creates a submatch that will contain zero or more word characters. The second element should look familiar to you but the last element is special. The "</" marks the beginning of a closing tag, but you can't just look for the first instance of this because there may be nested tags. Use the backslash to insert the first submatch (called a backreference) indicated by the number one. The desired text is in submatch 2, the second set of parentheses. You can test this by replacing "sarcasm" by another string.

Re: Regular Expressions Board

TCPlomp — Mon, 26 Jul 2010 07:21:16 GMT

It might be an idea to add a 'RegEx' group on the communities?

In this group we could host a page with a table showing 'input text', 'Regex', 'match' and a description of the actual working of the regex.

That way it's better to manage and search than a thread with numerous responses.

Ton

a Regular Expression to match a number

jcarmody — Fri, 30 Jul 2010 10:18:53 GMT

I was poking around for RegEx examples in the Community and found this one titled "Determining Whether a String Contains Only Numbers". I posted two alternate solutions and thought I'd post one of them here.

Read the expression like this: Look from the start of the string (^) for one or more digits ([\d]+) followed, optionally, by a decimal point (\.?) and zero or more digits ([\d]*) up to the end of the string ($) OR (|) look from the start of the string (^) for a decimal point followed by one or more digits ([\d]*) up to the end of the string ($).

This will find a number formatted like .#, #.#, #. or #

You can modify this to look for a number anywhere in your input string by removing the carets and dollar signs. You'll have to replace each instance of "\." if you use the wrong symbol to indicate the decimal point (ie: a comma). 😉

Re: one way to use a backreference

Ray.R — Mon, 16 Aug 2010 14:24:31 GMT

@jcarmody wrote:

Submatches are useful in extracting a part of your match but they also can also be reused in the regexp. An example is to match beginning and ending tags when you don't know what tag you're looking for.

The brackets are not special characters so they'll result in a literal match. The expression "(\w*)" creates a submatch that will contain zero or more word characters. The second element should look familiar to you but the last element is special. The "</" marks the beginning of a closing tag, but you can't just look for the first instance of this because there may be nested tags. Use the backslash to insert the first submatch (called a backreference) indicated by the number one. The desired text is in submatch 2, the second set of parentheses. You can test this by replacing "sarcasm" by another string.

This is a nice example that I can build from to write code that accesses webpages and extracts information from it.

I think I will exploit the power of RegExp's in the code. I may need some help along the way. At least I know where to post 😉

Re: one way to use a backreference

jcarmody — Mon, 16 Aug 2010 16:14:38 GMT

I did some more work on my answer to this post where the OP asked about "regular expression to match html tags". It may help with what you've proposed. I'm having some trouble with it but I'm hoping we can work them out as we go forward.

Re: one way to use a backreference

RobCole — Mon, 23 Aug 2010 19:45:18 GMT

I had to modify the expression for tags so I could work with multi-line strings. Like the following:

<parameter>Something</parameter>

</parameters>

</Action>

This is what I came up with to strip the tags off one level and skip the new line and tab characters. There is probably a more efficient way to do it, but this is working for me.

I left the input in / codes to show the tab and newline characters.

Rob

Re: one way to use a backreference

Ray.R — Mon, 13 Sep 2010 16:53:40 GMT

I'm still trying to wrap my head around even a simple regular expression in LabVIEW.

For instance, let's say I want to delete any instance of <div ****whatever is in here ****>

an example would be:

or even </div>

basically irradicating anything that contains "div" within < >.

How is that written in regEx ease?

~~<div R>~~

~~</div>~~

Re: one way to use a backreference

jcarmody — Mon, 13 Sep 2010 17:53:56 GMT

I'm not sure I understand what you're after, but here's my interpretation. You'll have to put this inside a loop to handle multiple search and replace operations.

The key here is to use a Character Class to search for the closing ">". Read the Search String as: "Find a '<' followed by any number of anything except a '>', followed by the '>', OR look for exactly '</div>'." The Search and Replace String function will do the rest (you'll need to right-click the node and select Regular Expression).

Everything inside the square brackets constitutes a Character Class. Beginning one with a caret immediately after the opening bracket tells the Regexp engine to negate the characters in this class. The pipe | is the OR operator; it matches either the preceeding or following expression.

Re: one way to use a backreference

Ray.R — Mon, 13 Sep 2010 18:23:57 GMT

Thanks Jim,

I can get one or the other to work in seperately:

<div[^>]*> works

and

</div> works (obviously),

but

<div[^>]*>|</div>

(<div[^>]*>)|(</div>)

didn't work.

is it misinterpreting the pipe as a character?

have a look at the output sting you posted.. 😉

Re: one way to use a backreference

jcarmody — Mon, 13 Sep 2010 18:40:49 GMT

Could you post some of the text you're working with?

Re: one way to use a backreference

Ray.R — Mon, 13 Sep 2010 19:57:56 GMT

Here's a section of the html code. It simply represents the <div> stuff that I want to remove.

As a matter of fact, I need to parse and extract some information from within the html code.

As I said... just a small portion and sample information to show the <div fields>. I do have something that works, but it is in 2 steps. I tried to get it working in a single step as the code you provided.

Obviously the code below is not intended to work, simply an example.. 😉

============================================

    <div id="post_message">


      <div class="usenext">
<div class="title">
<a rel="nofollow" href="http://website.com" target="_blank">some website</a>
</div>
Some miscellaneous message, text and things that are not important. <a href="http://www.somewebsite.com" title="a website somewhere on the net"><strong>Title of website</strong></a>.</a>
</div><br /><br /><div style="width:200px"><div align="center"><br />
<img style="max-width: 500px; cursor: pointer;" onclick="window.open(this.src)" src="http://anotherwebsite.com/image.png" border="0" alt="" /><br />

</div><br />
<br />
<br />
</div>

Re: one way to use a backreference

Ray.R — Mon, 13 Sep 2010 20:04:08 GMT

As a matter of fact, you could do it for the source code from this thread's source code... Unfortunately, if I try to post part of the source code from this thread, it quickly goes beyond the limited characters for the message..

There are plenty of <div fields> to remove... 🙂

The purpose of posting was to learn to improve my regEx knowledge.. I'm trying to figure out why the "OR" '|' didn't work....

Thanks.

Let's try this portion of the source code from this thread...

<div class="MinimumWidthContainer">
       <div class="min-width-wrapper">
           <div class="min-width">

                       <div class="lia-content">



    <div class="lia-quilt-layout-one-column lia-quilt-forum-topic-page lia-quilt">
   <div class="lia-quilt-row-header lia-quilt-row">

       <div class="lia-quilt-column-common-header lia-quilt-column-single lia-quilt-column-24 lia-quilt-column">
           <div class="lia-quilt-column-alley-single lia-quilt-column-alley">
   <div class="lia-component-quilt-header lia-quilt-layout-header lia-quilt-header lia-quilt">
   <div class="lia-quilt-row-title lia-quilt-row">
       <div class="lia-quilt-column-page-title lia-quilt-column-left lia-quilt-column-10 lia-quilt-column">
           <div class="lia-quilt-column-alley-left lia-quilt-column-alley">
   <h2 class="lia-component-common-widget-page-title PageTitle"><a class="lia-link-navigation" id="link_0" href="/t5/BreakPoint/bd-p/BreakPoint">BreakPoint</a></h2>
</div>
       </div><div class="lia-quilt-column-site-navigation lia-quilt-column-right lia-quilt-column-14 lia-quilt-column">
           <div class="lia-quilt-column-alley-right lia-quilt-column-alley">
   <div class="lia-component-common-widget-site-navigation SiteNavigationDropDown">

   <div class="lia-menu-navigation-wrapper" id="siteNavigationDropDown">
   <div class="lia-menu-navigation lia-js-click-menu">
       <div class="dropdown-default-item"><a name="title" class="lia-link-navigation default-menu-option lia-js-menu-opener" rel="nofollow" id="dropDownLink" href="#">Go To</a>
           <div class="dropdown-positioning">
               <div class="dropdown-positioning-static">

   <ul id="dropdownmenuitems" class="lia-menu-dropdown-items">

In the above example, all the <div class...> should dissappear..