Is it possible to distinguish between code, strings and comments by using regular expressions?
I'm trying to build a formatter for PLSQL-Code in C# and I can't manage to solve the problem, that strings, respectively quotes ( " or ' ) can also appear in comments, negating their special character and vice versa, comment indicators ( -- or /* ) appearing in strings, being nullified as a special character sequence.
Is there a solution, or are regular expressions simply the wrong way to handle this?
Best regards,
Philipp
You might find an idea how to reliably identify comments by studying a source code of Perl module Regexp::Common::comment. The module supports matching also PL/SQL comments.
Related
According to the received wisdom MS Word (more or less) supports find/replace with use of regular expressions. I have a simple regular expression:
^(C[[:alpha:]]*)(\d*)(.*)$
That I'm running on the data:
indSIMDdecile
CSdeccrim12006
CSdeccrim12006
CSdeccrim12009
CSdeccrim12009
CSdeccrim12012
CSdeccrim12012
CSdeceduc12004
CSdeceduc12004
CSdeceduc12006
CSdeceduc12006
CSdeceduc12009
CSdeceduc12009
CSdeceduc12012
CSdeceduc12012
CSdecemp12004.x
I'm interested in returning the first word prior to the digit 1, which works as demonstrated on regex101 here.
Problem
I would like to the same but in MS Word (v. 15.18 on Mac). After getting error messages of trying to supply unsuitable syntax I learned that MS Word does not support to the full regex syntax. I simplified my expression to something on the lines:
but the search does not find any strings and nothing gets replaced. Hence my questions, is it possible to use MS Word on Mac with regex?
The linked help website hints that something like that should be possible, but so far now luck.
The simple answer is "no", if you mean "Does Mac Word have a UI feature that lets you use one of the modern dialects of regex?" Word's Find/Replace only supports its own Regular Expression syntax.
In this case, I think the following will give you what you need:
Find with wildcards:
(C)([!1]#)(1)
and a replace by
\1
(If you also had to find "C1", then that doesn't work, and unfortunately nor does
(C)([!1]{0,})(1)
because Word does not allow 0 in the {,} pattern)
But there is a problem with "#". If the text the "#" is looking for is long, the find/replace may fail. There is supposed to be a 255 limit, but it seems rather more arbitrary than that. (I have long suspected a buffer overrun type error in the Word code, but perhaps there is a simpler explanation).
If you mean, "is there any way to use modern regex with Word?", then the answer is "Yes, but you only get to operate on a copy of the text in the document. You will need to create your own code to do the 'replace' part of the find replace, and that means that you would have to deal with any of the issues such as preserving formatting that Word's built-in find/replace might get right for you.
On the Windows side, people who want a better regex than Word's often use VBScript's regexp object because it is easily used from VBA. VBA itself only really has the "like" operator, which also only has fairly crude pattern matching abilities. I think there are examples of VBScript rexexp use on StackOverflow. On the Mac side, you would either have to use VBA and "shell out" to one of the built-in Mac/Unix utilities to do your finding (and perhaps replacing), or perhaps use Applescript or Javascript application scripting to do it. As far as I can remember Applescript does not have a 'modern' regex built-in either.
[As a bit of history, Word's "regular expressions" were I think introduced in Word 6, around 1993, at a time when most dialects of regex were much more crude than they are today. I don't think Word's version has moved along much at all - it probably added some Unicode support at some point, but that's probably about it. I assume that people using modern regex don't regard it as regex at all, and I personally prefer not to call Word's Regular Expressions 'regex' precisely for that reason.]
Hey stackoverflow community. Ive need help with huge information file. Is it possible with regular expression to find in this tag:
<category_name><![CDATA[Prekiniai ženklai>Adler|Kita buitinė technika>Buičiai naudingi prietaisai|Kita buitinė technika>Lygintuvai]]></category_name>
Somehow replace all the other data and leave only 'Adler' or 'Lygintuvai'. Im using Altova to edit xml files, so i cant find other way then find-replace. And im new in the regex stuff. So i thought maby you can help me.
#\<category_name\>.+?gt\;([\w]+?)\|.+?gt;([\w]+?)\]\]\>\<\/category_name\>#i
\1 - Adler
\2 - Lygintuvai
PHP
regex101.com
Fields may contain alphanumeric characters without spaces.
If you want to modify the scope of acceptable characters change [\w] to something other:
[a-z] - only letters
[0-9] - only digits
etc.
It's possible, but use of regular expressions to process XML will never be 100% correct (you can prove that using computer science theory), and it may also be very inefficient. For example, the solution given by Luk is incorrect because it doesn't allow whitespace in places where XML allows it. Much better to use XQuery or XSLT, both of which are designed for the job (and both work in Altova). You can then use XPath expressions to locate the element or attribute nodes you are interested in, and you can still use regular expressions (e.g. in the XPath replace() function) to process the content of text or attribute nodes.
Incidentally, your input is rather strange because it uses escape sequences like > within a CDATA section; but XML escape sequences are not recognized in a CDATA section.
Regular Expressions are incredible. I'm in my regex infancy so help solving the following would be greatly appreciated.
I have to search through a string to match for a P character that's not surrounded by operators, power or negative signs. I then have to insert a multiplication sign. Case examples are:
33+16*55P would become 33+16*55*P
2P would become 2*P
P( 33*sin(45) ) would become P*(33*sin(45))
I have written some regex that I think handles this although I don't know how using regex I can insert a character:
The reg is I've written is:
[^\^\+\-\/\*]?P+[^\^\+\-\/\*]
The language where the RegEx will be used is ActionScript 3.
A live example of the regex can be seen at:
http://www.regexr.com/39pkv
I would be massively grateful if someone could show me how I insert a multiplication sign in middle of the match ie P2, becomes P*2, 22.5P becomes 22.5P
ActionScript 3 has search, match and replace functions that all utilise regular expressions. I'm unsure how I'd use string.replace( expression, replaceText ) in this context.
Many thanks in advance
Welcome to the wonder (and inevitable frustration that will lead to tearing your hair out) that is regular expressions. You should probably read over the documentation on using regular expressions in ActionScript, as well as this similar question.
You'll need to combine RegExp.test() with the String.replace() function. I don't know ActionScript, so I don't know if it will work as is, but based on the documentation linked above, the below should be a good start for testing and getting an idea of what the form of your solution might look like. I think #Vall3y is right. To get the replace right, you'd want to first check for anything leading up to a P, then for anything after a P. So two functions is probably easier to get right without getting too fancy with the Regex:
private function multiplyBeforeP(str:String):String {
var pattern:RegExp = new RegExp("([^\^\+\-\/\*]?)P", "i");
return str.replace(pattern, "$1*P");
}
private function multiplyAfterP(str:String):String {
var pattern:RegExp = new RegExp("P([^\^\+\-\/\*])", "i");
return str.replace(pattern, "P*$1");
}
Regex is used to find patterns in strings. It cannot be used to manipulate them. You will need to use action script for that.
Many programming languages have a string.replace method that accepts a regex pattern. Since you have two cases (inserting after and before the P), a simple solution would be to split your regex into two ([^\^\+\-\/\*]?P+ and P+[^\^\+\-\/\*] for example, this might need adjustment), and switch each pattern with the matching string ("*P" and "P*")
I would like to remove hundreds on onmouseover events from my code. the evt all pass different variables and I want to be able to use dreamwaever to find and replace all the strings with nothing.
Here is an example
onmouseover="parent.mv_mapTipOver(evt,'Wilson');"
onmouseover="parent.mv_mapTipOver(evt,'Harris');"
onmouseover="parent.mv_mapTipOver(evt,'Walker');"
I want to run a search that will identify all of these and replace/remove them.
I have tried seemingly infinite permutations of things like:
onmouseover="parent.mv_mapTipOver(evt,'[^']');"
or
onmouseover="parent.mv_mapTipOver(evt,'[^']);"
or
onmouseover="parent.mv_mapTipOver(evt,[^']);"
or
onmouseover="parent.mv_mapTipOver(evt,'[^']+');"
And many more. I cannot find the regular expression that will work.
Any/all help would be appreciated.
Thanks a ton!
"." and "(" have special meaning in regular expressions, so you need to escape them:
onmouseover="parent\.mv_mapTipOver\(evt,'[^']+'\);"
I'm not sure if this is correct dreamweaver regex syntax, but this stuff is standard enough.
Try this one:
onmouseover="parent\.mv_mapTipOver\(evt,'.+?'\);"
And see it in action here.
When using reg expressions you have to be very careful about how you handle white space. For example the following piece of code will not get caught by most of the reg expressions mentioned so far because of the space after the comma and equals sign, despite the fact that it is most likely valid syntax in the language you are using.
onmouseover= "parent.mv_mapTipOver(evt, 'Walker');"
In order to create regexp that ignore white space you must insert /s* everywhere in the regexp that white space might occur.
The following regexp should work even if there is additional white space in your code.
onmouseover\s*=\s*"parent\.mv_mapTipOver\(\s*evt\s*,\s*'[A-Za-z]+'\s*\);"
I have an MFC project where I need to read and compare various configuration strings from (xml-)files.
The problem is that they could contain one or multiple special characters like STX, ETX, LF, CR ... and so on.
An idea is using regex. I could simply write the full regex pattern in the files and compare them with a match function.
As I looked this up via google and msdn, there were two different(?) regex frameworks for MFC but I don't see any difference between them nor do I see if they can solve my problem, meaning handle special characters.
Do any of you have an experience with those frameworks? Can you recommend one or can you think of another solution for this problem?
Many thanks in advance.
I recommend std::regex or boost::regex over non standard alternatives. Also, they are able to support special characters.