8 replies to this topic

[jasonalien]

Hello,
I need to find all of the same words in a text file. For example I need to find all "apples" in that file. After that I need to take the string which is next to the apple. I would appreciate any help, especially for the first part.

[gregwarner]

Scan character by character for the letter 'a'. Once found, check the next letter to see if it's a 'p'. If it is, check the next letter to see if it's a 'p'. If it is, check the next letter to see if it's an 'l'. And the next letter if it's an 'e'. And the next if if it's an 's'. If you successfully match each of these letters, you've matched the word "apples", so store the occurrence and reset for the next character and resume. If at any point during the middle of the word, if the next letter doesn't match the expected letter, drop out of the match, reset, and continue with comparing 'a' with the next character.

[jasonalien]

Thanks for the answer
How will I iterate in the text file? How will I store the occurence?

[gregwarner]

How will I iterate in the text file?

http://cplusplus.com...am/istream/get/
Use istream::get() to grab one character at a time. Place this in a while loop testing ios::good() to see if you can keep reading characters from the file.

How will I store the occurence?

Keep a counter of how many times the search word appears. Increment this on a successful match (i.e., when you match the last letter in the word.) Keep a state variable to tell where you are in the word as you are matching it.

Also, store the position in the file where the word occurs, if this is important. You'll need an array of integers for this. Index 0 will store the byte location of the first occurrence, index 1 will be the location of the 2nd occurrence, etc.

[BlackRabbit]

are you familiar with regular expressions library ? that will make it easy for you, capturing in groups of previous word - apple - next word, giving you a collection you can iterate ...

[gregwarner]

If you decide to go with BlackRabbit's suggestion, there are several regular expression libraries for C/C++. Two that I know of are PCRE (http://pcre.org/) and Boost (http://www.boost.org...html/index.html). Regex will certainly be a much quicker solution, however, if the purpose is to learn how to perform these matches algorithmically on your own, regex won't give you that.

[kernelcoder]

It seems you are a beginner? If yes, I think you better follow gregwarner's first. Then when you will know how it is working, you can go for BlackRabbit's suggestion for REGEX way (using regular expression libraries) to know about it.

[jasonalien]

I'm a "very" beginner. Thanks for helping..

[Avrillavigne]

It's a bit nonsensical to describe string-finding algorithms as O(N) or O(N*N) when there are really two parameters that determine the runtime, N_haystack and N_needle. Pretty much any algorithm is O(N_haystack) when N_needle = 1. Most are O(N_haystack * N_needle) or better, and you may be able to assume N_needle <= C
dapfor.com/en/net-suite/net-grid/tutorial/threadsafe-bindinglist