CODEWITHSUNDEEP
cfileextractstore
Mirel
Mirel

Reputation: 29

Store multiple types of data from file

Can someone help me with the following request:

I have a file with only one line that contains the following groups separated by a space

Here are some examples:

    0 0.3 -1 +3 0xc 'N'

    2.12211 1 -100 'N' 0xac 'N'

Basically the first line might contain real numbers,integer numbers,hexadecimal numbers, the character 'N' and each group is separated by a space

Is there any way to create a vector of multiple types(in C,not in C++) that will contain the elements separated by spaces(also the hexadecimal numbers should be converted to decimal numbers?

For the first example it shall be 

 V[0]=0
 V[1]=0.3
 V[2]=-1
 V[3]=3
 V[4]=12
 v[5]='N'

For the second example it shall be 

V[0]=2.12211   
V[1]=1
V[2]=-100
V[3]=3
V[4]='N'
v[5]=172
V[6]='N'

Upvotes: 1

Views: 266

Answers (2)

Adam Trhon
Adam Trhon

Reputation: 2935

Rule #1: Avoid premature optimization.

Define a type of your data:

enum data_type {
    data_type_integer = 0,
    data_type_float,
    data_type_char,
    ...
};

Now define structure to contain these data:

struct data {
    enum data_type data_type;
    int data_integer;
    float data_float;
    char data_char;
};

Of course this wastes some space, but unless you write a db engine or something, remember rule #1. If it really bothers you, pack the data (not data_type) into union. But if you make a mistake with union, it will bite you (as the compiler cannot check that, for example, you stored char and read int).

Now you have a single type to store various data types.

To parse it, there are various approaches. I suggest to read each item as a string and then parse the string using sscanf. The reason for this two-step approach is that you can check the string contents to determine the type: If it contains 'N', it's 'N'. If it contains 'e', 'E' or '.' it is float, etc.

Upvotes: 2

Code-Apprentice
Code-Apprentice

Reputation: 83537

Is there any way to create a vector of multiple types(in C,not in C++)

This seems like a good use of a union. You can define it as

union data {
    float f;
    int i;
    char n;
};

Now declare an array of unions:

union data V[20];

And populate it like so:

 V[0].i=0
 V[1].f=0.3
 V[2].i=-1
 V[3].i=3
 V[4].i=12
 v[5].n='N'

In the parsing logic, you will need to decide which field of the union to use in order to ensure the correct value is stored as you intend.

Upvotes: 1

Related Questions