I did 3 experiments

1. rb mode
2. r mode
3. r mode with seek of different offset bytes

with open("small_file.txt", "w") as f:
    f.seek(2)
    f.write("Content")

with open("small_file.txt", "rb") as f:
    content = f.read()
    print(f"START|{content}|END")

f.seek(2) during write is to insert some nullbytes (2 being arbitrary).

Output: START|b'x00x00Content'|END (No surprise here)

If i change rb to r, i get START|Content|END.

Question 1: What are the underlying concepts here to explain the behaviour of not reading (or reading but not printing?) null bytes, is it specified in any documentation?

Then i added a seek before reading

with open("small_file.txt", "w") as f:
    f.seek(2)
    f.write("Content")

with open("small_file.txt", "r") as f:
    f.seek(3)
    content = f.read()
    print(f"START|{content}|END")

which returns START|ontent|END.

I see the behaviour is anything <=2 bytes in f.seek during read will produce START|Content|END with no truncation of data.

I understand this threshold is directly related to f.seek(2) used during write, meaning if write used f.seek(3), the threshold during read will be 3.

Question 2: Why no truncation when <=2? Is this explainable by the same answer to question 1?

I also tried seeking beyond the end in both rb and r using f.seek(9) (9 because 2 offset during write + 7 letters in content) and got START||END.
This is expected and explained by https://docs.python.org/3.8/tutorial/inputoutput.html#methods-of-file-objects . It never mentioned whether this empty string applies to rb or r mode so i assume both, please correct if wrong.

If the end of the file has been reached, f.read() will return an empty string (”)