QRegExp - does not return all printable characters.

Anonymous_Banned275

QRegExp rx("^\p[ -~]*$");
The above code does not work - in C++. .
Not sure about the syntax.

This code works fine
QRegExp rx("(\d+)(\s*)(cm|inch(es)?)");

I also cannot find meaning of "\p"in QT doc.
.

SGaist

Hi,

QRegExp has been deprecated in Qt 5 and removed in Qt 6.

Please use QRegularExpression.

You can use the QRegularExpression Example to design and validate your regex.

Anonymous_Banned275

@SGaist Success !

This simple code "matches" first word (#Waiting.. ) in 'line" string , but LAST character.
Time to read more and add stuff...
Thanks

 QRegularExpression re("[ -~] ");
            QRegularExpressionMatch match = re.match(line);
            qDebug() <<"Regulsr expression ...test match " ;
            if (match.hasMatch()) {
                qDebug() <<"Regulsr expression ...match.hasMatch()" ;
                QString matched = match.captured(0); 
                qDebug()<< "Matched " << matched; 
                // ...
            }
            else
            {
                qDebug()<< "NO MATCH  ";
            } ...

JonB

@AnneRanch said in QRegExp - does not return all printable characters.:

QRegularExpression re("[ -~] ");

This does not match a word. ~~It matches a space/hyphen/tilde followed by a space.~~ [EDIT No, see next answer below.] So match.captured(0) will return precisely 2 characters. [EDIT This is still true.]

QRegExp rx("^\p[ -~]*$");
QRegExp rx("(\d+)(\s*)(cm|inch(es)?)");

Both of these --- and it would be same for QRegularExpression --- would require the \s in the literal strings to be written as \\ for each one, as per C++ rules for literal string characters.

Anonymous_Banned275

@JonB I cannot argue this - the symbols inside [ ] were posted on another forum and advertised as " match all printable characters ". Obviously do not.
Unfortunately the match is "character space " - as you said. So I get last character of the first world and a space - which is little hard to actually "see".

JonB

@AnneRanch said in QRegExp - does not return all printable characters.:

I cannot argue this - the symbols inside [ ] were posted on another forum and advertised as " match all printable characters ".

And I cannot argue with this! Let me be the first to admit when I have made mistake :) I read too quickly, and quite forgot that - (hyphen) is the one character with "special significance" when appearing inside the [...] one-of-characters entity. There (and only there) it represents a "character range", from the character before it (space here) to the character after it (tilde here), which does indeed match all printable (ASCII) characters.

So re("[ -~] ") does indeed match "any printable character followed by a space". That still does mean precisely 2 characters rather than a whole word, which would have to be the last character of a "word" followed by a space. However, it will also match 2 spaces, which is presumably not desired.

There are several ways to pick out a "word" if that is what you want to do instead. The simplest would probably be (in your C++ code):

QRegularExpression re("\\w+");

but it all depends "which" word you might want in a line, what to do about the other characters in the line, etc.

KroMignon

@AnneRanch said in QRegExp - does not return all printable characters.:

I cannot argue this - the symbols inside [ ] were posted on another forum and advertised as " match all printable characters ". Obviously do not.
Unfortunately the match is "character space " - as you said. So I get last character of the first world and a space - which is little hard to actually "see".

With QRegExp, the dash symbol in [] is used to define a range of characters (cf. Qt documentation https://doc.qt.io/qt-5/qregexp.html#sets-of-characters).

I think the right way would be to use backslash as escape sequence:

RegularExpression re("[ \\-~] ");

JonB

This post is deleted!

KroMignon

@JonB said in QRegExp - does not return all printable characters.:

I don't think that's right: I believe that would match a space or any character between \ and ~!

I have to admit that I don't have understand what is the goal of this regex.
My comprehension was that he want to match only on 3 characters space, bash and tilde.
But maybe I am wrong and the goal was to get any character between space and tilde (booth included) followed by a space.
So the regex should be RegularExpression re("[ -~] ");
But this also don't made sense to me.

JonB

@KroMignon
Hi Kro. I had to delete my response you have quoted after I had written it! In my day, it used to be the case that to place a literal - inside a [...] character range you had to put it as the first or last character, so that it could not be interpreted as a "range" operator, which is what I was writing up. However, it seems that since those dim, dark UNIX System V.0 days you can now escape a literal - via \- so you can have it anywhere in the characters, just as you wrote. Hence I deleted my response to your response!

But maybe I am wrong and the goal was to get any character between space and tilde (booth included) followed by a space.

That is exactly what the OP must have got from the source she took it from. However as pointed out that would include space-followed-by-space, which is unlikely to be desirable/robust. And at best it would have picked out the last character of a word followed by a space, which still seems an "unusual" thing to want. I have since suggested \w+ as probably the most canonical way of picking out "a word".

Anonymous_Banned275

I hope this may help this discussion.
I am positing the "raw" file - all three lines.
I have added a silly loop to see if I can mach more than FIRST word / character in each line .

Stream file 
Stream file  "Waiting to connect to bluetoothd...\u001B[0;94m[bluetooth]\u001B[0m#                                                                              \u001B[0;94m[bluetooth]\u001B[0m#                         Agent registered"
Regulsr expression ...test match 
Regulsr expression ...match.hasMatch()
Matched  "Waiting"
Matched  ""
Matched  ""
Matched  ""
Matched  ""
linea:  "Waiting to connect to bluetoothd...\u001B[0;94m[bluetooth]\u001B[0m#                                                                              \u001B[0;94m[bluetooth]\u001B[0m#                         Agent registered"
Stream file 
Stream file  "\u001B[0;94m[bluetooth]\u001B[0m# "
Regulsr expression ...test match 
Regulsr expression ...match.hasMatch()
Matched  "0"
Matched  ""
Matched  ""
Matched  ""
Matched  ""
linea:  "\u001B[0;94m[bluetooth]\u001B[0m# "
Overload QString
 File bt_utility_library.cpp 
function  ProcessCommand 
@ line   162

Again this is code under construction and posted in hope to help.

I have used
QRegularExpression re("\w+"); //matches first word / character

Also to clarify
at this point of learning using QRegulaExpression I just what to extract ALL words from the FIRST line.

I'll tackle deleting the control characters AKA match only printable characters later.

Anonymous_Banned275

Guess what I found - nice interactive tool

https://regex101.com/

JonB

@AnneRanch said in QRegExp - does not return all printable characters.:

QRegularExpression re("\w+"); //matches first word / character

Exactly as I wrote in response to your code earlier, if this is what you have written in your C++ source code it will not work.

I explained about the \ in C/C++ source code and said you would need to have:

QRegularExpression re("\\w+");

Is that what you have?

at this point of learning using QRegulaExpression I just what to extract ALL words from the FIRST line.

If you want to match/extract multiple matches you must change to calling QRegularExpression::globalmatch(). This is described in https://doc.qt.io/qt-5/qregularexpression.html#global-matching, and the example there is exactly what you are asking for:

Global matching is useful to find all the occurrences of a given regular expression inside a subject string. Suppose that we want to extract all the words from a given string, where a word is a substring matching the pattern \w+.

QRegularExpression re("(\\w+)");
QRegularExpressionMatchIterator i = re.globalMatch("the quick fox");

QStringList words;
while (i.hasNext()) {
    QRegularExpressionMatch match = i.next();
    QString word = match.captured(1);
    words << word;
}
// words contains "the", "quick", "fox"

Anonymous_Banned275

Update / conclusion (??)
The attached TEST code works as expected using the ORIGINAL "range" pattern. Using "global match " solved it. It matches EACH individual character in the string - not the entire word.
Note there are NO "escape characters" in the pattern.
It also "prints" control characters - which is NOT desirable - so I will work on that.
It is probably slow, but it does not matter in my app.

#ifdef BYPASS
              terminal outoput 
                bluetoothctl
                Agent registered
                [bluetooth]#

               string to analyze / match 
                linea:  "Waiting to connect to bluetoothd...\u001B[0;94m[bluetooth]\u001B[0m#                                                                              \u001B[0;94m[bluetooth]\u001B[0m#                         Agent registered"
                linea:  "\u001B[0;94m[bluetooth]\u001B[0m# "
#endif

 //               QRegularExpression re("(\\w+)");  works OK 
                QRegularExpression re("([ -z])");       original range 
                QString pattern = re.pattern(); // pattern == "a third pattern"
                qDebug()<<"Pattern function " << pattern;
                TRACE_TextEdit->addItem("TRACE pattern ");
                TRACE_TextEdit->addItem(pattern);
                QRegularExpressionMatchIterator i = re.globalMatch(line);
                QString text = "Match ";
                QStringList words;
                while (i.hasNext()) {
                    QRegularExpressionMatch match = i.next();
                    QString word = match.captured(1);
                    text += word;
                    TRACE_TextEdit->addItem(text );
                    words << word;
                    qDebug()<<" Global match (words)  " << words;
                }

JonB

@AnneRanch said in QRegExp - does not return all printable characters.:

It matches EACH individual character in the string - not the entire word.

That is because you have chosen to use re("([ -z])") instead of the re("(\\w+)") suggested. It is the + symbol which causes match whole word/multiple characters instead of per character. Did you really want to match each individual character separately? Up to you.