Why Doesn't the Loop Repeat? - c

Question

My question is this: Given a sentence, delete only the last "THE" from that sentence if it starts with the word "WE", otherwise delete all "THE"s.
my code does not delete second THE after the first THE.
What's wrong?
int main(){
char sent[100];
int i;
printf("Enter a sentence: ");
gets(sent);
if(sent[0] == 'w' && sent[1] == 'e'){
for (i=2;i<strlen(sent);i++){
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' '){
while( i < strlen(sent)){
sent[i] = sent[i+1];
sent[i] = sent[i+2];
sent[i] = sent[i+3];
sent[i] = sent[i+4];
i++;
}
printf("\n%s",sent);
}
}
}
else
for (i=0;i<strlen(sent);i++){
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' '){
while( i < strlen(sent)){
sent[i] = sent[i+1];
sent[i] = sent[i+2];
sent[i] = sent[i+3];
sent[i] = sent[i+4];
i++;
}
printf("\n%s",sent);
}
}
return 0;
}

Answer 1

Your problem is the inner loop to copy the remaining text:
for (i=2;i<strlen(sent);i++){
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' '){
while( i < strlen(sent)){
sent[i] = sent[i+1];
sent[i] = sent[i+2];
sent[i] = sent[i+3];
sent[i] = sent[i+4];
i++;
}
printf("\n%s",sent);
}
}
There are 2 main issues:
If you remove a word you must continue to search at the same location. After the loop iteration is done, the i++ is executed, advancing to the next position. Therefore you need to negate that effect by adding a i--; in case you removed an occurance of the.
Your inner loop to copy the remaining text messes up your counter for the outer loop. After you copied the text, i points to the last char terminating your loop. You must use a separate counter.
... and some more issues:
Minor issue:
You have useless assignments in the copy loop. If you assign a series of values to sent[i] only the last one will have an effect afterwards.
Minor issue:
Using strlen as loop condition causes lot of extra runtime. Only use it if necessary. For the inner loop simply check the length once and store it.
gets is evil. You should use fgets instead.
An updated version could look like this:
size_t len = strlen(sent); // Calculate limit only once
for (i=2; i < len; i++) {
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' ') {
size_t j = i;
while (j <= len-4) { // also copy the terminating 0 byte
sent[j] = sent[j+4];
j++;
}
printf("\n%s",sent);
len -= 4; // String was shortened
i--; // Check current position again.
}
}
Note:
You should also take into account that the last word could be the without an extra space...
You might also think about using standard C library functions like strstr...

Answer 2

The error that terminates your loop on the first occurrence of the word the is that the inner while loop modifies the loop variable i of the outer for loop.
for (i=0;i<strlen(sent);i++){
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' '){
/* after this loop i points to the end of the string */
while( i < strlen(sent)){
sent[i] = sent[i+1];
sent[i] = sent[i+2];
sent[i] = sent[i+3];
sent[i] = sent[i+4];
i++;
}
printf("\n%s",sent);
}
}
So when the while loop has moved i to the end of the string, the outer for loop will terminate as well.
This problem can be fixed by using a separate variable for the inner loop.
I also tried to fix the out-of-bounds access and moved strlen out of the loop. (still incomplete, see below)
/* the length is constant except when we remove "the " */
size_t len = strlen(sent);
for (i=0; i < len; i++){
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' '){
/* use separate variable for inner loop */
int j = i;
/* limit len-3 tp prevent sent[j+4] from reading past the NUL ('\0') character */
while( j < len-3){
/* useless assignments to the same array element removed */
sent[j] = sent[j+4];
j++;
}
len -= 4; /* we removed 4 characters */
printf("\n%s",sent);
}
}
This code has still the problem that it doesn't remove "the the " because the for loop won't check the same position again after removing "the ".
One option would be to decrement i after removing "the ". If the string starts with "the the " this would temporarily set i to -1. It will be incremented again by the for statement.
I prefer to increment i only if "the " was not found.
That's why I move the i++ from the for statement into the loop body.
/* the length is constant except when we remove "the " */
int len = strlen(sent);
for (i=0; i < len; /* no i++ here */){
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' '){
/* use separate variable for inner loop */
int j = i;
/* limit len-3 tp prevent sent[j+4] from reading past the NUL ('\0') character */
while( j < len-3){
/* useless assignments to the same array element removed */
sent[j] = sent[j+4];
j++;
}
len -= 4; /* we removed 4 characters */
printf("\n%s",sent);
} else {
/* advance to the next character only if we didn't just remove "the " */
i++;
}
}
The while loop can also be written as a for loop.
/* the length is constant except when we remove "the " */
int len = strlen(sent);
for (i=0; i < len; /* no i++ here */){
if(sent[i] == 't' && sent[i+1] == 'h' && sent[i+2] == 'e' && sent[i+3] == ' '){
/* use separate variable for inner loop */
/* limit len-3 tp prevent sent[j+4] from reading past the NUL ('\0') character */
for(int j = i; j < len-3; j++){
/* useless assignments to the same array element removed */
sent[j] = sent[j+4];
}
len -= 4; /* we removed 4 characters */
printf("\n%s",sent);
} else {
/* advance to the next character only if we didn't just remove "the " */
i++;
}
}
Note: You can write the loops in various different ways, so there are other valid solutions, as you can see from the comments.