How many times a word appears in a text?

[Net2022]

Hello, I have googled and searched the forum, but I can't find anything to get my code right. I want to search a .txt file and see how many times a keyword appears on demand. I've tried using a loop with the match pattern function, but I'm not getting the correct results. I appreciate any help given.

[crossrulz · Accepted solution]

You need to wire up the shift register to the "start index" input on the Match Pattern.

Personally, I would just use the Search And Replace String, setting the "replace all" to TRUE and then just use the "Number Of Replacements" output.

[RTSLVU]

Try this: 

Honestly, crossrulz solution above is much better.

[paul_cardinale]

One unpleasant complication is figuring out what delimits a word.

[RTSLVU]

---

@paul_cardinale wrote:

One unpleasant complication is figuring out what delimits a word.

---

Irrelevant, the program searches for an exact match and returns how many time that exact match was found.

[altenbach]

---

@paul_cardinale wrote:

One unpleasant complication is figuring out what delimits a word.

---

I agree that the problem is much more complicated, because there are plenty of delimiters (space, tab, period, linefeed, double-quote, numbers, etc.) so it is important to only use consecutive strings from the correct lexical class. Just searching for a substring will give you plenty of false positives (e.g. searching for "trust" would also count the substring in "trustworthy" as a "word", etc.). You also need to ignore upper/lowercase.

Still, this is a simple homework problem to solve. Try it!

Some comments to the original code:

• No, it is not OK to spin the outer loop millions of times per second, redlining one CPU core and heating your computer.
• NO, it is not OK to place a wait inside the inner loop. Searching should progress as fast as possible (seconds, not hours for long files!)
• Please don't maximize the front panel and diagram to the screen! (during debugging you need to see both, as well as the help windows!)
• The indicator you call "frequency" has the wrong representation and does not display a frequency at all.
• Your inner loop always finds the same word forever and will never stop unless the word exists zero times.
• Since the keyword control should be constant during the inner loop, it belongs outside it.
• Why do you have a shift register is you don't even look at its output?
• etc.

[paul_cardinale]

---

@RTSLVU wrote:

---

@paul_cardinale wrote:

One unpleasant complication is figuring out what delimits a word.

---

Irrelevant, the program searches for an exact match and returns how many time that exact match was found.

---

Certainly not irrelevant.  Suppose the text contains " forum ", and you search for the word "rum" without regard for delimiters; you would get a match there in "forum".  But that is not the word "rum" and shouldn't be counted as such.

[paul_cardinale]

---

@altenbach wrote:

---

@paul_cardinale wrote:

One unpleasant complication is figuring out what delimits a word.

---

I agree that the problem is much more complicated, because there are plenty of delimiters (space, tab, period, linefeed, double-quote, numbers, etc.) so it is important to only use consecutive strings from the correct lexical class. Just searching for a substring will give you plenty of false positives (e.g. searching for "trust" would also count the substring in "trustworthy" as a "word", etc.).

Still, this is a simple homework problem to solve. Try it!

---

And you have to make sure that words at the beginning and end don't get missed for lack of a delimiter on one side.

[RTSLVU]

---

@paul_cardinale wrote:

---

@RTSLVU wrote:

---

@paul_cardinale wrote:

One unpleasant complication is figuring out what delimits a word.

---

Irrelevant, the program searches for an exact match and returns how many time that exact match was found.

---

Certainly not irrelevant.  Suppose the text contains " forum ", and you search for the word "rum" without regard for delimiters; you would get a match there in "forum".  But that is not the word "rum" and shouldn't be counted as such.

---

Oh, good point... I hadn't thought about it that much.

[altenbach]

See if this can give you some ideas.

It would be even easier to place all words and their count into a map (try it!).

NOTE that you should set the keyword control to "limit to single line", else a linefeed might sneak in. You might actually want to verify that only a single word is entered (not shown)!