HELP! picking words from a file and copying it into a new file

Someone please help me.
I need to write a java program that takes the contents of a html file (ignores the tags) and copies the individual words into a separate master file with the name of the html file that it came from by each word. When i repeat this with another file, it shud put those words into the same mater file, but no overwrite the previous data i.e.
test.html
< H1 > This is a Heading < H1 >
masterfile.txt
This test.html
is test.heml
a test.html
heading test.html

What have you done so far on this program?

You might want to read up on RegEx to help you bypassing those tags.

Bob,
Welcome to Javaranch.
This is a nice parsing problem Bob.
There are many things you will need to consider such as what exactly is a word and other unwanted contents such as with the SCRIPT tag which you probably don't need. You will also need to avoid duplicates from the masterfile.txt and a whole bunch of other things.
I think you should use the power of regex to strip the HTML tags. Read in variable <code>HTMLtext</code> a paragraph or even the whole file. Start with the simple pattern <.*?> to strip the tags (using Sting.replaceAll). Then split the words using \s+.
I have tested this approach with Perl and it sounds like a good starting point:

I ran this in cygwin and it seems to work fine. There is a lot in this Perl one liner but it follows the same logic which I outlined. Combined with the shell and the '>>' we do a 'sort -u' for uniqueness and append to the masterfile.txt
I will just explain the perl bit
<code>perl</code> the command
<code>-0</code> (That's zero) defines the record separator usually an octal value is supplied. In this case nothing is supplied which means there is no rec sep and the whole file will be slurped.
<code>-n</code> tells perl to loop implicitly on the input. (which is supplied by '*.html')
<code>-e</code> takes a perl statement as an argument.
Then starts the single quote to protect the perl snippet from shell interpretation.
<code>s</code> substitute (In java you'd say <code>replace</code>
<code>/</code> Delimiter to begin the pattern
<code><.*?></code> The pattern to search for. A '<' followed by any character .*? (non-greedy)and a closing '>'.
<code>//</code> The closing delimiter and the replacement (which is nothing)
<code>s</code> The /s lets '.' to match the newline character. (in Java you would <code>compile</code> with <code>Pattern.MULTILINE</code>
<code>g</code> The /g modifier stands for replace globally, that is, all occurrences. (in Java use <code>replaceAll</code>
<code>print</code> simple I hope. (in Java you do <code>System.out.print</code>

<code>map {"$_ $ARGV\n"}</code> evaluates it's argument (split) and creates a string in the { BLOCK } as word, followed by $ARGV which Perl assigns the 'current' file in progress
<code>split</code> The argument to map which splits the input based on the default regex (\s+)
Then there is the closing quote followed by the rest of the command.
Cheers,
Leslie

HTML is not my strength, why doesn't the <code>code</code> tag work all the time

First of all, <code> and </code> aren't even valid HTML tags. Secondly, HTML is not enabled on this board, according to the message next to the text area where we type messages. However, if you replace < > with [ ], this will use the UBB tags, which will mark the code similarly to what you expect.
Layne
[ April 10, 2003: Message edited by: Layne Lund ]

Thanks Layne for the clarification. I have never tried this but I guess UBB tags will work inline.
Let's see... (If there would only be a way where I can see the post before posting

)

So there you go it dosen't. So how would I get this idea?

Leslie,
If you click the little pen and paper icon on any of your own posts, then you can edit them.

Just put all the code between the [ code ] and [ /code ]. This will put it between lines like in your previous post. You should only have one set of these tags surrounding a single block of code. Individual lines of code can just be on their own lines in between the tags.