Map member of class produces error - why?

I have a map member of a class.
Inside a function of the class, I call nameOfMap.insert(), and the compiler gives an error saying:

Quote:

---

error: expected unqualified-id before '.' token

---

You should understand that we can't tell anything unless you show the code.

Quote:

---

Originally Posted by Elysia

You should understand that we can't tell anything unless you show the code.

---

Code:

---

#include <map>
#include <vector>
#include <string>

using namespace std;

class CaClass {

  public:
    CaClass(); //constructor
    void doStuff();

  private:
    typedef map<unsigned int, string> aMap;
    vector<string> aStringVector;
    vector<unsigned int> aUnsignedIntVector;

};

CaClass::CaClass() {

  //nothing in here that deals with the map....

}

void CaClass::doStuff() {

  unsigned int currentNum = aUnsignedIntVector.at(0);
  unsigned int i = 0;
  do {
        aMap.insert(aMap::value_type(currentNum - 1, aStringVector.at(currentNum - 1))); //note that I'm using currentNum - 1 here, since every element of the unsigned int vector is 1 more than the corresponding element in the string vector
        currentNum = aUnsignedIntVector.at(i+1);
        i++;
  while(i < aUnsignedIntVector.size());

}

---

Where are the template arguments for map?
std::map<Key, Value>
Also, you could easily write this...

Code:

---

aMap.insert(aMap::value_type(currentNum - 1, aStringVector.at(currentNum - 1)));

---

...as this:

Code:

---

aMap[currentNum - 1] = aStringVector.at(currentNum - 1);

---

Quote:

---

Originally Posted by Elysia

Where are the template arguments for map?
std::map<Key, Value>

---

Yeah, I noticed that before you posted.
My bad. I already fixed it in my last post.
I meant to make the map declaration in the class look like this:

Code:

---

typedef map<unsigned int, string> aMap;

---

Because, that's how it is in the real code.

Quote:

---

Also, you could easily write this...

Code:

---

aMap.insert(aMap::value_type(currentNum - 1, aStringVector.at(currentNum - 1)));

---

...as this:

Code:

---

aMap[currentNum - 1] = aStringVector.at(currentNum - 1);

---

---

Thanks, I will try that.

Quote:

---

Code:

---

aMap.insert(aMap::value_type(currentNum - 1, aStringVector.at(currentNum - 1)));

---

---

O_o

According to the code, `aMap' is a type; it is not an instance. Is that really your code?

Soma

Ok, using your example changed the error message to this instead:

Quote:

---

error: expected unqualified-id before '[' token

---

Quote:

---

Originally Posted by phantomotap

O_o

According to the code, `aMap' is a type; it is not an instance. Is that really your code?

Soma

---

Yes, because I'm copying the style of someone else (actually 'Bubba' on these forums) when writing a map because I don't know much about maps. I originally tried the code with the map declaration looking like this instead:

Code:

---

map <unsigned int, string> aMap;

---

But then the compiler printed out a message saying something like the map wasn't a class type, so I changed it to use a typedef. Anyway, I just changed it back to that after reading your post, and now its getting past that error, so it appears that it is solved.

When you encounter syntax errors, you must post the code or we cannot help you.
Using a typedef or not doesn't matter. If it works without the typedef and not with it, then you're doing something wrong.

Quote:

---

Originally Posted by Elysia

When you encounter syntax errors, you must post the code or we cannot help you.
Using a typedef or not doesn't matter. If it works without the typedef and not with it, then you're doing something wrong.

---

Well, apparently it did matter using the typedef in this case, since once I changed the insert calls in my code to using the '[]' and '=' operators instead, like in your example, the compiler still gave an error message on the map line, when I was still using typedef. Then, following phantom's post, I removed the 'typedef' before the map declaration inside the class, and now the compiler doesn't give me any error on the same line of code adding stuff to the map.

Like I said, if it works without the typedef but not with it, then you have an error in your code.

Quote:

---

Originally Posted by Elysia

Like I said, if it works without the typedef but not with it, then you have an error in your code.

---

Maybe. I've been wondering if I need to call the constructor of the map object inside my class constructor first before I try calling a member function of map.
For some reason, though, I thought an object of map would automatically be created by the compiler when you use the '.' operator to access a member function of my member map.

Well here's the problem.

aMap is either a type, in which stuff like this is valid:

aMap::value_type (foo);

Or it is an instance, which lets you insert things:

aMap[bar] = foo;

It cannot be both.

Quote:

---

Originally Posted by Programmer_P

Maybe. I've been wondering if I need to call the constructor of the map object inside my class constructor first before I try calling a member function of map.
For some reason, though, I thought an object of map would automatically be created by the compiler when you use the '.' operator to access a member function of my member map.

---

The constructor of the map is invoked before the constructor body of your class.
The map is ready to use since it's an object inside your class (not a pointer nor a reference).

Quote:

---

Originally Posted by whiteflags

Well here's the problem.

aMap is either a type, in which stuff like this is valid:

aMap::value_type (foo);

Or it is an instance, which lets you insert things:

aMap[bar] = foo;

It cannot be both.

---

And yet it would not let me use the insert() function to insert things into the map, when it was just an instance and not a typedef. It was only when I changed it to use the '[]' and '=' operators, like in your example, that it got past the compiler error. Why is that?

Quote:

---

Originally Posted by Programmer_P

And yet it would not let me use the insert() function to insert things into the map, when it was just an instance and not a typedef. It was only when I changed it to use the '[]' and '=' operators, like in your example, that it compiled. Why is that?

---

Show the code in which you tried to "use the insert() function to insert things into the map, when it was just an instance and not a typedef".

Quote:

---

Originally Posted by Elysia

The constructor of the map is invoked before the constructor body of your class.
The map is ready to use since it's an object inside your class (not a pointer nor a reference).

---

Ok, then I have no idea why I'm now getting another error with a different member map, one that is a copy of the original map and is assigned with the '=' operator of the map class inside a member function.

Show your code, please.

Quote:

---

Originally Posted by laserlight

Show the code in which you tried to "use the inshttp://cboard.cprogramming.com/newreply.php?do=newreply&p=954825ert() function to insert things into the map, when it was just an instance and not a typedef".

---

EDIT: Delete reply...

My code has compiled. :p

Quote:

---

Originally Posted by Elysia

Show your code, please.

---

No need. It was a noobie mistake. :p I accidentally forgot to prefix the member function definition that assigns one map to another map to make a copy of the map, then returns the copy, with the class name followed by the "::" operator, which is why it was saying the map copy name didn't exist in the scope.

All good now.

Quote:

---

aMap[currentNum - 1] = aStringVector.at(currentNum - 1);

---

Map::Insert is a far better choice even though it is syntactically more complex.

Depends on what you need. If you are just going to insert without regard as to whether it's there or not, then the index operator might do the job.
Any other reason it shouldn't be preferred?