Extract most occurred word in a text field

[ibobo]

Hello, 

I m trying to find 5 most occurred word in a text field.

for exemple  text : "Hello John how are you ? Jonh may I ask you to go out to night? I like going out at night, you remember."

so the result will be : You (x3), John (x2), Out (x2), Night (2x) Hello (1) 

 

could some one help me with that 

 

Thank you

[Lee Smith]

iBobo,

Not sure why you need this, but if you have a list of the words to count? You might be able to use the PatternCount for this? 

There is a Custom Function I located that could also be helpful uniqueWords

Lee

pattern.fmp12

[ibobo]

 Thanks Lee Smith

I need because I want to generate key words from a description.

For exemple i m selling a printer, the word printer will be repeated few time in the description, so the script will extract few most repeated word to make them key words from that description. And one of the key word will be "printer" and other most repeated.

I m not sure if I m clear with my ideas 

[ibobo]

I think I m getting close, I change a bit the "pattern.fmp12"

I add a new table with two fields "keyword" and "occurence"

so now instead of answer "John: came up 1 times"  > "John" go to field (keyword) and "1" go to field "occurence", then I ll tri them by bigger occurence and get top 5 word occurred in a text :)))) and then generate list of keywords ))) 

I remarked some issu with "the" I should keep the "the" with the word instead of making it as a different word. 

 

Ibobo  

 

pattern.fmp12

[Lee Smith]

Hi ibobo,

My Bad, I thought this was a file I made a long time ago for a similar need. However,  the credit for this file goes to rwoods who posted it in response in this thread. Pattern count/Keyword density probelm,

I must have deleted my file or I'm not remember what I named it. It used a calculation and not a script, and I provided the words to count.

I'll continue to look or will recreate it for you.

Lee

[Fitch]

I took it a little further. Fun exercise. I didn't comment my script but if you stop and think about each step you'll see what's going on. The key is to keep track of the top count as you go. Also FYI I left the word list field in there so you could watch it while debugging but I'd probably move that to a variable, the field isn't really needed. Let me know if you have any questions.

pattern.fmp12

[ibobo]

Hi Lee, Hi Fitch  

 

thanks a lot to helping me ! , that work's perfectly )) 

there is one adjustment,  i don't know what is the best ? to ignore the word "for", "and" , "the", "to" or make it as one word ? so it will count occurence of combined word as one word ?

 

Edited December 14, 2017 by ibobo

[Fitch]

I think using Substitute to ignore them is fine, unless you really want to count them -- then I'd probably put them in their own $wordList "bucket" (or put them in the $count = 1 bucket).

[ibobo]

Thanks a lot, I ll do as you said :))) 

Have a nice week end