Confusion About Array and Strings in C
What is the difference between S1
, S2
and S3
?
char S1[6];
S1[0] = 'A';
S1[1] = 'r';
S1[2] = 'r';
S1[3] = 'a';
S1[4] = 'y';
char S2[6] = {'A','r','r','a','y'};
string S3 = "Array";
When I run the program using if (strcmp(a,b) == 0)
where a, b = S1, S2, S3
. He shows that S2
they are the S3
same, but S1
also S2
different. Why is this so ?? Why aren't all three equivalent?
And when I add back '\0'
to the S1b
, S1c
. All 3 are the same. This is clear.
BUT why in my first test S2
and are the S3
same then? I didn't include it either '\0'
. And I suspect that S1
u S2
should be the same, but not S2
both S3
.
Can anyone tell me why my thought is wrong?
Thank you for your responses. I tried and changed the settings like this:
char S1[5];
S1[0] = 'A';
S1[1] = 'r';
S1[2] = 'r';
S1[3] = 'a';
S1[4] = 'y';
char S2[5] = {'A','r','r','a','y'};
string S3 = "Array";
And now it is clear that S2
they S3
do not match, since they differ by '\0'
. However, I'm still a little confused as to why S1
and S2
don't match this time if I use strcmp to compare the two?
Compare the actual memory values in the arrays:
- S1 has 6 elements large, but you only specify values for 0-5, the 6th element is not explicitly set, so it retains whatever value is in memory prior to allocation.
- S2 is similar to S1, only 5 elements are provided, however
{,}
any additional elements are cleared when using the syntax . So, itchar foo[5] = { 1, 2 }
is identicalchar foo[5] = { 1, 2, 0, 0, 0}
. - S3 uses a string syntactic way of initializing an array, which creates an array
char
(orwchar_t
) with an optional element set to\0
(null terminator).
Visually:
S1 = 0x41, 0x72, 0x72, 0x61, 0x79, 0x??
S2 = 0x41, 0x72, 0x72, 0x61, 0x79, 0x00
S3 = 0x41, 0x72, 0x72, 0x61, 0x79, 0x00
Note that strcmp
you are facing a security issue with the problem: it does not have a length parameter, it continues searching until it encounters \0
, which may never be (i.e. until a segfault or violation access). Use a safer type function strncmp
or (if using C ++) type instead std::string
.
It shows that S2 and S3 are the same, and S1 and S2 is different.
S3 contains a null terminator that S1 does not. This string S3 = "Array";
means
| A | r | r | a | y | \0 |
Bye S2
| A | r | r | a | y | \0 |
While S1
| A | r | r | a | y | Garbage |
Comparing S1 and S2 could result in UB (I suppose) because S1 is not nul-terminated and there is no length we pass to strcmp
.
#include <stdio.h>
#include <string.h>
int main(void)
{
char S1[6];
S1[0] = 'A';
S1[1] = 'r';
S1[2] = 'r';
S1[3] = 'a';
S1[4] = 'y';
S1[5] = 0;
char S2[6] = {'A','r','r','a','y', 0};
printf("%d" ,strcmp(S1,S2));
return 0;
}
Outputs:
0
strcmp () starts comparing the first character of each string. If they are equal to each other, it continues the next pairs until the characters are different or until a terminating null character is reached.
I don't think it's safe to compare S1 and S2 using this. The input to strcmp is the address of the first character. S1 does not end with zero. Although 6 bytes are allocated in both cases, S1 [5] is not initialized. Most likely they have some garbage value. The risk here is that strcmp will end up comparing unallocated memory also looking for a diff or null character. It can even lead to a crash or access error.
Visualizing memory alignment S1, S2, S3 could be something like this
S1 = A | r | r | a | y | ?
S2 = A | r | r | a | y | 0
S3 = A | r | r | a | y | 0
Any comparison between S2 and S3 is safe. S1 vs S2 or S3 may not exist.
Just an addition to existing answers
char S2[6] = {'A','r','r','a','y'};
string S3 = "Array";
Both are NULL-terminated and hence strcmp()
works well and says they are both the same. Although S1
assignment is explicit, there is no NULL termination for this array. Thus, it is not a valid string in C. Therefore, use strcmp()
can lead to undefined behavior.
Point c S3
is that S3 is a read-only string literal. Basically, these values are stored in read-only locations. So when you try to write something in S3
after initialization, you might see a crash. Therefore, we must keep this in mind when using assignments likeS3