Remove “b” and “ac” from a given string

Given a string, eliminate all “b” and “ac” in the string, you have to replace them in-place, and you are only allowed to iterate over the string once. (Source Google Interview Question)

Examples:

acbac   ==>  ""
aaac    ==>  aa
ababac  ==>   aa
bbbbd   ==>   d

The two conditions are:
1. Filtering of all ‘b’ and ‘ac’ should be in single pass
2. No extra space allowed.

The approach is to use two index variables i and j. We move forward in string using ‘i’ and add characters using index j except ‘b’ and ‘ac’. The trick here is how to track ‘a’ before ‘c’. An interesting approach is to use a two state machine. The state is maintained to TWO when previous character is ‘a’, otherwise state is ONE.
1) If state is ONE, then do NOT copy the current character to output if one of the following conditions is true
…a) Current character is ‘b’ (We need to remove ‘b’)
…b) Current character is ‘a’ (Next character may be ‘c’)
2) If state is TWO and current character is not ‘c’, we first need to make sure that we copy the previous character ‘a’. Then we check the current character, if current character is not ‘b’ and not ‘a’, then we copy it to output.

// A C++ program to remove "b" and 'ac' from input string #include <iostream> using namespace std; #define ONE 1 #define TWO 2   // The main function that removes occurences of "a" and "bc" in input string void stringFilter(char *str) {     // state is initially ONE (The previous character is not a)     int state = ONE;       // i and j are index variables, i is used to read next character of input     // string, j is used for indexes of output string (modified input string)     int j = 0;       // Process all characters of input string one by one     for (int i = 0; str[i] != '\0'; i++)     {         /* If state is ONE, then do NOT copy the current character to output if            one of the following conditions is true            ...a) Current character is 'b' (We need to remove 'b')            ...b) Current character is 'a' (Next character may be 'c') */         if (state == ONE && str[i] != 'a' && str[i] != 'b')         {             str[j] = str[i];             j++;         }           // If state is TWO and current character is not 'c' (otherwise         // we ignore both previous and current characters)         if (state == TWO && str[i] != 'c')         {             // First copy the previous 'a'             str[j] = 'a';             j++;               // Then copy the current character if it is not 'a' and 'b'             if (str[i] != 'a' && str[i] != 'b')             {                 str[j] = str[i];                 j++;             }         }           // Change state according to current character         state = (str[i] == 'a')? TWO: ONE;     }       // If last character was 'a', copy it to output     if (state == TWO)     {         str[j] = 'a';         j++;     }       // Set the string terminator     str[j] = '\0'; }   /* Driver program to check above functions */ int main() {     char str1[] = "ad";     stringFilter(str1);     cout << str1 << endl;       char str2[] = "acbac";     stringFilter(str2);     cout << str2 << endl;       char str3[] = "aaac";     stringFilter(str3);     cout << str3 << endl;       char str4[] = "react";     stringFilter(str4);     cout << str4 << endl;       char str5[] = "aa";     stringFilter(str5);     cout << str5 << endl;       char str6[] = "ababaac";     stringFilter(str6);     cout << str6 << endl;       return 0; }

Output:

ad

aa
ret
aa
aaa

An extension of above problem where we don’t want “ac” in output at all:
The above code looks fine and seems to handle all cases, but what if input string is “aacacc”, the above code produces output as “ac” which looks correct as it removes consecutive occurrences of ‘a’ and ‘c’. What if the requirement is to not have an “ac” in output string at all. Can we modify the above program to produce output as empty string for input “aacacc” and produce output as “d” when input is “abcaaccd”? It turns out that it can also be done with given restrictions. The idea is simple. We need to add following lines inside for loop of the above program.

if (j>1 && str[j-2] == 'a' && str[j-1] =='c')   j = j-2;

See this for different test cases of modified program.

A Simpler Solution to the Original Problem:

// // A C++ program to remove "b" and 'ac' from input string #include<bits/stdc++.h> using namespace std;   void stringFilter(char *str) {     int n = strlen(str);       int i = -1;  // previous character     int j = 0;   // current character       while (j < n)     {         /* check if current and next character forms ac */         if (j < n-1 && str[j] == 'a' && str[j+1] == 'c')             j += 2;           /* if current character is b */         else if (str[j] == 'b')             j++;           /* if current char is 'c && last char in output            is 'a' so delete both */         else if (i >= 0 && str[j] == 'c' && str[i] == 'a')             i--,j++;           /* else copy curr char to output string */         else             str[++i] = str[j++];     }     str[++i] = '\0'; }   // Driver program to test above function int main() {     char str1[] = "ad";     cout << "Input => " << str1 << "\nOutput => ";     stringFilter(str1);     cout << str1 << endl << endl;       char str2[] = "acbac";     cout << "Input => " << str2 << "\nOutput => ";     stringFilter(str2);     cout << str2 << endl << endl;       char str3[] = "aaac";     cout << "Input => " << str3 << "\nOutput => ";     stringFilter(str3);     cout << str3 << endl << endl;       char str4[] = "react";     cout << "Input => " << str4 << "\nOutput => ";     stringFilter(str4);     cout << str4 << endl << endl;       char str5[] = "aa";     cout << "Input => " << str5 << "\nOutput => ";     stringFilter(str5);     cout << str5 << endl << endl;       char str6[] = "ababaac";     cout << "Input => " << str6 << "\nOutput => ";     stringFilter(str6);     cout << str6 << endl << endl;       char str[] = "abc";     cout << "Input => " << str << "\nOutput => ";     stringFilter(str);     cout << str << endl << endl;     return 0; }

Output:

Input => ad
Output => ad

Input => acbac
Output =>

Input => aaac
Output => aa

Input => react
Output => ret

Input => aa
Output => aa

Input => ababaac
Output => aaa

Input => abc
Output =>

Thanks to Gaurav Ahirwar for suggesting this simpler solution.

This article is contributed by Varun Jain. Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above