Array of Pipe: Will parent process load the array fast enough?

Question

    for (int i = 0; i < number_processes; i++){

        while (read(fps[i][0], &each_record_read, sizeof(struct rec)) > 0){
            if (records_container[i] == NULL){
                records_container[i] = &each_record_read;
            }
        } // I want to do something to records_container here.
    }

I have a program that its purpose is for the parent process to read data from each pipe that connects to several children.

I have define an array called records_container that contains each data type that I am going to read from the pipe, and I want to do something to this array after the while loop.

My question is this: when parent is loading data to records_container this array, is parent loading speed fast enough so that when I want to do something to records_container, I can always make sure all data is already being loaded perfectly?

Let's assume records_container is {0, 0, 0, 0} in this case (I know I want to load struct rec, but this doesn't matter really), the ideal situation that I want it to happen is that I want to deal with the loaded array, which is {1, 2, 3, 4} in this case, will parent read each number fast enough (which is 1, 2, 3, 4 from each pipe in this case) so that I won't deal with something like {1, 2, 0, 0}, or {1, 0, 0, 0} ... that kind stuff?

Answer 1

This does not work.

while (read(fps[i][0], &each_record_read, sizeof(struct rec)) > 0)

is wrong and you can get a short read. When dealing with a structured pipe, we typically use something like this:

ssize_t read_block(int source, void *buffer, ssize_t len)
{
    char *work = buffer;
    while (len) {
        ssize_t delta = read(source, work, len);
        if (delta < 0) return -1; /* ERROR */
        if (delta == 0) return 0; /* EOF */
        work += delta;
        len -= delta;
    }
}

/* ... */

    while (read_block(fps[i][0], &each_record_read, sizeof(struct rec)) > 0){

Now it works.