Count number of words in a text?

[bleh]

Hello

The title says it basically: How can i count the number of words in a text?

[Melba23]

bleh,

Just one of the threads that appeared when I searched the forum.

M23

P.S. The "Search" facility is at top right.

[water]

What do you mean by "text"? A string with space separated words, a Word document ...?

[bleh]

Sorry, Melba23.

 

What do you mean by "text"? A string with space separated words, a Word document ...?

It can be a string or a text file or whatever. The text is a normal text with punctuation etc. Just like a Wikipedia entry for example.

[water]

Search the forum for one of the word count examples as Melba suggested. Then decide which characters should denote a new word (space, hyphen, @CRLF, @LF ...).

[UEZ]

I don't know how accurate this code is:

$sText = "GDI + allows you to easily manipulate 2D, without having to select from the dc, the pen, the font, so the brush can restore the DC to its original state before returning (it's hard enough with GDI). We can manipulate images, scale, rotate, translate, shear, or mix these functions quite easily. The pen has several functions to make indents. You can define preset tips for lines such as the tips of arrows or creating custom tips. The brushes are quite numerous and diverse, we can make very easily degraded by passing the color of departure and arrival."
MsgBox(0, "Test", $sText & @CRLF & @CRLF & "Number of words: " & CountWords($sText))

Func CountWords($sText)
    Local $aResult = StringRegExp($sText, "(\w+)", 3)
    If @error Then Return SetError(1, 0, 0)
    Return UBound($aResult)
EndFunc

Br,

UEZ

[water]

According to the help file of StringRegExp when using "(w+)" a word is a consecutive number of this characters: a-z, A-Z, 0-9 or underscore (_)

[UEZ]

Are these 3 words: " * * *."?

What is the definition of a word?

Br,

UEZ

[water]

Let's see how the OP defines "word".

[mikell]

Are these 3 words: " * * *."?

As there is neither a space nor an apostrophe I would say one...

What about  StringRegExp($sText, "[^s']+", 3) ?

Edit

assuming there is no typo or grammatical mistake in the text

Edited July 4, 2013 by mikell

[UEZ]

@mikell: it has spaces (underscore)-> "_*_*_*."

Br,

UEZ

[jchd]

We didn't have a chance to see what "word" means to the OP. OTOH current PCRE implementation is compiled without PCRE_UCP, sadly (yes I do heavily insist on that). As a bad consequence, the easy way, b, can't be used in the general case.

Hopefully PCRE is kind enough to provide us with Unicode-wide h and v (and their negation, resp. H and V) to match horizontal and vertical "spaces" but the problem now shifts to detecting punctuation, which would require UCP support to be of general use.

So once again we're stuck with half-backed solutions which work for (?i)[a-z] English only despite AutoIt claiming to support Unicode (which it doesn't really).

Edited July 4, 2013 by jchd

[bleh]

Let's see how the OP defines "word".

 

"A single distinct meaningful element of speech or writing, used with others (or sometimes alone) to form a sentence"

I don't know whats unclear. I want to count the number of words in texts using AutoIt.

For example: The quote above are seven words. The punctuation (" and .) would not count as a word.

The link of Melba23 seems to work fine, btw.

Thanks to you all.

 

Edited July 5, 2013 by bleh

[orbs]

how about StringSplit with the delimiters string being a combination of common separators,

like space & @TAB  & @CR & @CRLF & @LF and whatever.

[jchd]

"A single distinct meaningful element of speech or writing, used with others (or sometimes alone) to form a sentence"

I don't know whats unclear. I want to count the number of words in texts using AutoIt.

If that's the definition you're going to use, then you're in deeper problem than you seem to think.

Please apply your own word count definition to the following translations of the very same (modulo Google translate accuracy!) sentence, all of them matching your definition.

"This is a well-formed sentence that meets the definition." (english)

"Voilà une phrase bien formée qui répond à la définition." (french)

"這是一個結構完整的句子，符合定義。" (traditional chinese)

"זהו משפט בנוי היטב שעונה על ההגדרה." (hebrew)

"นี่คือประโยคที่ดีขึ้นที่ตรงตามคำนิยาม" (thaï)

"これは、定義を満たす整形文です。" (japanese)

"இந்த வரையறையை சந்திக்கிறது என்று ஒரு நன்கு வடிவமைக்கப்பட்ட சொற்றொடர் உள்ளது." (tamul)

... and so many others I won't bother to list, but you get the idea.

[bleh]

Seriously? Why pick it apart? What's the point?

Here you go:

I want to count the words of formatted, punctuated and well written texts based on the latin script.

What's the "Latin script"? http://en.wikipedia.org/wiki/Latin_script

What's a "word"? https://en.wikipedia.org/wiki/Word

What's "punctuation"? http://en.wikipedia.org/wiki/Punctuatin

"What does a text with words look like?" Like this: One, two and three.

"But how many words would that be?" 4.

[Melba23]

bleh,

 

Why pick it apart? What's the point?

Because we get so many people who ask a question and then complain about proposed solutions becasue they do not meet the special cases that they omitted to mention at the beginning.

Thank you for explaining clearly what it you are asking. Does the thread to which I linked you in post #2 not do what you need?

M23

[bleh]

Does the thread to which I linked you in post #2 not do what you need?

 

From my ealier reply (#13):

The link of Melba23 seems to work fine, btw.

Thanks to you all.

Edited July 5, 2013 by bleh

[jchd]

Well, if by latin script you intend a-z and A-Z, then yes. But your ironic answer is the first time where "latin" was used.

So much for precision.

Please recognize that your previous wordings were overly ambiguous. I wasn't picky, just trying to read your mind at a distance.

This forum receives posts from worldwide users so unless a precise context is clear, answers should be as general as possible.

[bleh]

Well, maybe you should read before you reply. Since everyone with common sense could understand what i meant with words and text after reading what i wrote. Especially the example in #13 and even more so after i wrote that Melbas example worked.

Please recognize that it's also save to assume that i would have said it if i wanted to count words in traditional Chinese, Egyptian hieroglyphs or Martian.

Edited July 5, 2013 by bleh