Return position of first character of a particular word

[gmeredith17]

I would like to search the contents of a variable for a particular word and return the character position of its first letter. I've had a look in the help file but couldn't find a ready made function. Before I embark on creating my own I wanted to check that there isn't a standard function that I have failed to find.

The background for this is that I would like to replace whole sections of text in a number of text files and replace them with new text.

So if there is a standard function to help with this can someone point me in the right direction.

Thanks

Geoff

[gmeredith17]

Re: Return position of first character of a particular word

AT2() might be the answer.

[G Gabriel]

Re: Return position of first character of a particular word

The background for this is that I would like to replace whole sections of text in a number of text files and replace them with new text.

Many different ways. It all depends. Is the text in question a certain word? a certain paragraph? some characters in a word?

[gmeredith17]

Re: Return position of first character of a particular word

I think AT2() will do it. I'm replacing one or more paragraphs. Each paragraph has a unique identifier for the start and finish.

I imagine the structure of the UDF being something like this;

Load text from file in variable vtext using .to_string()
Get total number of chrs using len() store in variable vtc.
Search for first word (unique identifier) and get position using at2() and store in variable pstart.
From this position search for end identifier using at2() and store in variable pend.
Have two more variables, vtext_start and vtext_end.
vtext_start will have the contents of vtext from beginning to the pstart-1.
vtext_end will have the text contents of vtext from pend+1 to vtc.

Then the new text file content will be
New_text_file=vtext_start+vnew_text+vtext_end

Where vnew_text will be a variable containing the replacement text.

Finally using file.from_string() create a new text file.

I'm sure there are better or more efficient ways but this is my initial idea. Once I have got it working for one text file I will need to expand the function to perform it on a batch of files.

[Al Buchholz]

Re: Return position of first character of a particular word

Geoff

Sounds like you are going down a path that will lead you to string.scanner

[gmeredith17]

Re: Return position of first character of a particular word

Thanks Al. I will take some time to digest the options. I kind of skipped over this function in the help file index. Seeing the term scanner I must have subconsciously thought it was for use with scanning hardware. At first glance it likes like it has some powerful functionality.

[csda1]

Re: Return position of first character of a particular word

Hi Al,

Sounds like you are going down a path that will lead you to string.scanner

Bad choice IMHO. This is a class with methods (functions) that, as far as I know, Alpha never uses internally. It is not very fast in almost all circumstances compared to alternatives, and I suspect it was relegated to the past after Alpha installed RegEx functions.

If not using Regex functions (the best here), then I probably would use wordat() or wordatc() to find the word #, then compute the length using word() to get the (1st words + the delimiter) prior to that

for a regex function use something like this

textstring="This is a test string that embedtestembed has a test string more than once"
' This finds any string
?regex_grep(textstring,regex_escape("test"),"$(OFFSET)"+crlf(),"I")
= 11
33
49
' This finds only strings separated by a regex defined word bondrary
?regex_grep(textstring,"\b"+regex_escape("test")+"\b","$(OFFSET)"+crlf(),"I")
= 11
49
' This takes the 1st word in the list of found values and returns the offset as a number
?val(word(regex_grep(textstring,"\b"+regex_escape("test")+"\b","$(OFFSET)"+crlf(),"I"),1,crlf()))
= 11

[gmeredith17]

Re: Return position of first character of a particular word

Ira - thanks for the alternative.

It would seem Alpha added the string.scanner function after regex looking at the supported version from the help file. V6 for regex_grep and V7 for sring.scanner.

Would they not have used the regex within the string.scanner functions?

They state this in the help file.

A comparison with OCCURSI() showed that a WHILE...END WHILE loop using <StringScanner>.SkipToString()(similar to the above script) was approximately 60 times faster.

It seems they did something to improve the speed. Do you know the background to the functions and what they did get the improvements?

[csda1]

Re: Return position of first character of a particular word

Hi Geoff,

It would seem Alpha added the string.scanner function after regex looking at the supported version from the help file. V6 for regex_grep and V7 for sring.scanner.

Would they not have used the regex within the string.scanner functions?

When they referenced
A comparison with OCCURSI() showed that a WHILE...END WHILE loop

All that means is that a single line of code is faster than a loop, which is almost always a truism. Whether it is as fast as other methods that are essentially 1 line of low-level ("C" code types) functions, I'd kind of doubt it.

Stringscanner and regex functions are both available in V6, although there is an indication that regex was present in A5v5, but not exposed at the xbasic level.

I don't have any idea if they used regex in stringscanner, but it's probably code that was different than regex, as they have different structures.

It's been over a year or two since I last ran checks of stringscanner speeds, but if it were really useful, I'd suspect Alpha would have used it internally, which I don't see any indication that they do.

However, feel free to test it's speed. I think you will find it slower than regex, or any other low-level functions option.

[G Gabriel]

Re: Return position of first character of a particular word

Each paragraph has a unique identifier for the start and finish.

It is possible that you do not need either.
Can you provide a sample?

[gmeredith17]

Re: Return position of first character of a particular word

G,

I want to edit several html pages. For example, change all the footers or menu structure in each page. I think any of the suggestions that have been made will be fine and the speed is probably a moot point when dealing with relatively small text files. I suspect disc access will be the major overhead.

Ira, thank you for the extra information. I hadn't come across the regex functions so that looks like an area I will have to put some time into to.

Geoff

[Clunes]

Re: Return position of first character of a particular word

Try looking up word() in help - Say you wanted the P in Pearson in a variable containing Mark Pearson. I would use

Left(word(variable,2," "),1)

What you need to do in your case is find the word number

[CALocklin]

Re: Return position of first character of a particular word

FYI: Just some related info for anyone reading this thread.

Note that these two functions:

at() - case sensitive
atc() - not case sensitive

are similar to at2() except for the third argument. The at2() function takes a third argument which is a starting point while the third argument in the at() and atc() functions is the "occurrence number". So
atc( "the", "to the other of them", 2 )
would return 9 because the second "the" is found in "other".

If you really wanted to find just a specific word - which might be at the very beginning or at the very end - you can add some spaces to both the string you are searching for and the string you are searching through like this:
test_str = "The list will be sent to the other of them."
atc( " the ", " " + test_str + " ", 2 )

NOTE: The wordat() and wordatc() functions only find the first matching word. They cannot be used to find the nth occurrence of a word. Also, they return the word number - not the position number. If parsing a string, you will probably have to convert the word number to a position number - which is why I just use atc().

Edit: Note also that the at2() function is case sensitive and there doesn't appear to be a case insensitive alternative. This is the main reason I seldom use this function.

[csda1]

Re: Return position of first character of a particular word

Hi Cal,

Edit: Note also that the at2() function is case sensitive and there doesn't appear to be a case insensitive alternative. This is the main reason I seldom use this function.

For something that is not returning a string derived from the parameter inputs, in this case position of the string, you can make it case insensitive simply by uppercasing all strings, as in

AT2(upper(searchstring),upper(stringtosearch),startingposition)