Reading Binary file into std::vector<bool>

Question

Hello I am trying to write 8 bits from std::vector to binary file and read them back . Writing works fine , have checked with binary editor and all values are correct , but once I try to read I got bad data . Data that i am writing :

11000111 //bits

Data that i got from reading:

11111111 //bits

Read function :

std::vector<bool> Read()
{
    std::vector<bool> map;
    std::ifstream fin("test.bin", std::ios::binary);
    int size = 8 / 8.0f;
    char * buffer = new char[size];
    fin.read(buffer, size);
    fin.close();
    for (int i = 0; i < size; i++)
    {
        for (int id = 0; id < 8; id++)
        {
            map.emplace_back(buffer[i] << id);
        }
    }
    delete[] buffer;
    return map;
}

Write function(just so you guys know more whats going on)

void Write(std::vector<bool>& map) 
{
    std::ofstream fout("test.bin", std::ios::binary);
    char byte = 0;
    int byte_index = 0;
    for (size_t i = 0; i < map.size(); i++)
    {
        if (map[i]) 
        {
            byte |= (1 << byte_index);
        }
        byte_index++;
        if (byte_index > 7)
        {
            byte_index = 0;
            fout.write(&byte, sizeof(byte));
        }
    }
    fout.close();
}

Not sure why you're using a fixed-size character buffer, especially one that's...a single byte long? Maybe? Are you expecting the file to contain 7.3 bits in the future? — tadman
– tadman, Commented Jan 25, 2017 at 20:48
vector <bool> is not exactly the greatest thing to use for any purpose — user2100815
– user2100815, Commented Jan 25, 2017 at 20:48
a "std::vector" with the variable name "map" is basically asking your coworkers to hate you — RyanP
– RyanP, Commented Jan 25, 2017 at 20:49
There's literally no reason to divide by a float. Also assumptions like #2 are how you get cripplingly bad overflow bugs. You can design around those assumptions, but always handle exceptional cases. — tadman
– tadman, Commented Jan 25, 2017 at 20:53

Kerrek SB · Accepted Answer · 2017-01-25 20:56:36Z

3

Your code spreads out one byte (the value of buffer[i], where i is always 0) over 8 bools. Since you only read one byte, which happens to be non-zero, you now end up with 8 trues (since any non-zero integer converts to true).

Instead of spreading one value out, you probably want to split one value into its constituent bits:

for (int id = 0; id < 8; id++)
{
    map.emplace_back((static_cast<unsigned char>(buffer[i]) & (1U << id)) >> id);
}

edited Jan 25, 2017 at 20:56

answered Jan 25, 2017 at 20:49

Kerrek SB

480k96 gold badges904 silver badges1.1k bronze badges

Sign up to request clarification or add additional context in comments.

9 Comments

Egi Dijus Over a year ago

thats because i pack 8 bits into one byte when was writing data to file

Kerrek SB Over a year ago

@EgiDijus: So you also need to split your one byte back out into eight bits. I added some code.

Egi Dijus Over a year ago

thats what i am trying to do when reading into loops buffer[i] << id , bit shifting byte to get 8 bits

Kerrek SB Over a year ago

@EgiDijus: Shifting isn't enough. You also need to mask out the unwanted bits.

Cameron Over a year ago

Kerrek is right. Aside from the unsigned issues (never shift signed values), if you take 3 and shift it left by 1 you get a true value, shift it by 2 you still get a true value, etc. even though only the lower two bits are set.

|

Collectives™ on Stack Overflow

Reading Binary file into std::vector<bool>

1 Answer 1

9 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

9 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related