Files and Buffers

Warning

This page is still Work-in-Progress.

Files and Buffers

What is a File?

One of the most essential concepts in a data persistence library is data files. Files are how data propagates across sessions: when the program first runs, the wanted data is written to a file, which continues to exist even after the program closes. Then, when the program is then opened again, it may read from that file to restore its data.

What is a Buffer and Why Use It?

As you might know, directly writing to a file on every write() will quickly prove to be inefficient, because writing data to a disk is far slower than, say, directly writing it to memory. So many persistence libraries, including cpersist use write buffers. A write buffer is a section in memory data is written to before it is written to the disk. Let us demonstrate.

Example

#include <cpersist/cpersist.h>

int main() {
    auto file = cpersist::File("myfile");

    file.write("number", 3);
    file.commit();
}

Explanation

First, we create a cpersist::File object. Because of the reasons we discussed earlier, this does not immediately create a file. Instead, it creates a buffer that will later be written into a file. We call it "myfile" and not, say, "mybuffer" because, although it is indeed a buffer, it's often easier to think of it as a file.

file.write("number", 3); writes the number 3 under the key "number." We will discuss write() functions in a later page.

file.commit(); is the important part. This line takes the file's buffer and dumps it into a file with the same name. After this function is called, a file with the name myfile.bin has been created and written the number 3 to. If you try opening the file, however, you will not see the number 3 directly. This is because of the way information is serialized. We will discuss serialization in a future chapter.

---

← Installation Guide · Writing and Committing →