I am using a switch statement to determine if a value is a Float or a string, then compairing it to another value that is of the same type. It will not do the comparison for some reason.

For example:

swith(fieldType){
case FLOAT: switch (compareOp){
case LESS: if (*((float*)(int offset) < (*((float *)char *valPtr))) return TRUE; else FALSE#: break;
case STRING: switch (compareOp){
case LESS: if (((char*)(int offset)) < (valPtr)) return TRUE; else FALSE#: break;


Is the problem in the <? or in the derefrencing? or is is something else?

What language is that? It looks a lot like C, but not enough like C to be correct.

Some specifics:

What did you intend by the things I marked in red? Looks like you intended some kind of cast. But that is not the syntax for a cast.

Then the part I marked in purple, I don't even have a guess what it was supposed to mean.

Once you get the syntax correct, you also should know that comparing char pointers does not compare the strings. It compares the POINTERS.

C uses a char* like a string, but C doesn't really have a string type. You need various helper functions to do basic string operations, such as strcmp for comparing them.

You also seem to be assuming an int is the same size as a pointer. That is a bad idea. It will work in 32-bit x86, but will fail in many architectures including x86_64. You should learn correct use of void*

[QUOTE=vwinnecke;4929731]

swith(fieldType){
case FLOAT: switch (compareOp){
case LESS: if (*((float*)(int offset) < (*((float *)char *valPtr))) return TRUE; else FALSE#: break;
case STRING: switch (compareOp){
case LESS: if (((char*)(int offset)) < (valPtr)) return TRUE; else FALSE#: break;
QUOTE]

It is just sort of an algorithm. Did you see my questions at the bottom?

I will rewrite with some corrections:

For example:

switch(fieldType){
case FLOAT: switch (compareOp){
case LESS: if (*((float*)(int offset) < (*((float *)char *valPtr))) return TRUE; else FALSE; break;
case STRING: switch (compareOp){
case LESS: if (((char*)(int offset)) < (valPtr)) return TRUE; else FALSE; break;

(int offset)

This part is defined as an integer, so it was

int offset;

then in the code offset. I just put int offset so I wouldn't have to define it earlier for this example.

Yes. But no direct answer would be meaningful. You posted garbage code and asked too specific a question about which detail was wrong.

Not enough corrections.

Maybe it would help if you knew this is comparing values in a database?

I think what you are trying to do is something like:

If you're comparing values in a database, then I suggest that you use SQL queries to let the database-engine do the comparison. Let it return to you only the values that you are looking for.

Also, remember that floating-point values can't be meaningfully compared for 'equality.' You need to calculate the absolute-value of the difference between the two numbers and see if it is less than some small number. This can be expressed in SQL, e.g.:
(COL1 IS NOT NULL) AND (COL2 IS NOT NULL) AND (ABS(COL1 - COL2) < 0.0001)

Even if you see two numbers that are, say, "3.14" when printed, one might actually be "3.1415926" and the other might be "3.14159271" in their full, float-binary representation. (And in a database, they could also be decimal fields.)

You can use a ROUND() function to ask that the two values be rounded to an appropriate number of decimal places for subsequent comparison. These results might be directly comparable.

In the case of a database, remember NULL: the absence of any value at all. You can only compare numbers that exist.