alternative reading binary in python instead of C++

Question

I have a binary file and C++ code which can read that binary file like follow.

int NumberOfWord;
FILE *f = fopen("../data/vec.bin", "rb");
fscanf(f, "%d", &NumberOfWord);
cout << NumberOfWord< <endl;

This output is:

114042

I want to reimplement like above code in python.

with open("../data/vec.bin","rb") as f:
 b = f.read(8)
 print struct.unpack("d",b)[0]

but this code is not working. my output is:

8.45476330511e-53

My question are:

1) why integer has 8 byte in C++.

I never knows %d means double. but, actually the variable has a type of integer, but normally we output using "%d" in C++. It is weird.

2) How do I extract a real number in python

I want to extract a real number like above C++ code in python code. How do I that??

maybe, I misunderstand about struct module in python.

• Robert Jørgensgaard Engdahl: Nebrils answer will do. I'd like to add that the decimal format specifier in Python is also "d", e.g. printf("%d", 42); in C corresponds to print("{0:d}" % 42) in Python.
• Serge Ballesta: Am I drunk or fscanf(..., "%d", ) is for reading text file???
• hiroyuki tanaka: thanks guys. fscanf(..., "%d", ) is for reading binary file in this time.

Answer 1

As you have been able to read the file correctly with this C++ (or rather C) line, fscanf(f, "%d", &NumberOfWord);, I assume that your file contains a text representation of 114042. So it contains the bytes

0x31 0x31 0x34 0x30 0x34 0x32 ... or '1', '1', '4', '0', '4', '2', ...

When you open it in a text editor, you can see one single line 114042.

Now when you try to read if as binary with i format, you use the 4 first bytes of the file and actually get int('31313034', 16): 825308208. I could not reproduce what you get with d format for decoding it as double because I could not guess what comes in your file after the last digit...

If the number is alone on first line, it is easy: just read one line and convert it to an int:

with open("../data/vec.bin","rb") as f:
 print int(f.readline())

If there are other characters after the last digit, you will have to first use a regex (do not forget to import re) to get the numeric value and then convert it to an int:

with open("../data/vec.bin","rb") as f:
 line = f.readline()
 m = re.match(t'\s*\d*', line)
 print(int(m.group(0)))

TL/DR: Do not try to read a text file as if it contained a binary representation

• hiroyuki tanaka: Thanks for polite answer. I understand what the problem is. so, maybe data file is weird because that contained text and binary.
• Serge Ballesta: Mixing binary and text io is much harder in Python than in C because you have no mean to obtain the file position after a textual read: Python library has read a full buffer and the file pointer is generally far beyond the end of line. You should (IMHO in another question) explain the structure of the file that you want to read,

Answer 2

In C format strings, %d is short for decimal.

In Python, d is short for double.

If it's an integer, you should use i in struct.unpack call.

with open("../data/vec.bin","rb") as f:
 b = f.read()
 print struct.unpack("i",b)[0]

• hiroyuki tanaka: Thanks! I don't know that difference. but, I ran using "i" like your code. but my output in python is 808726833. It's a different from C++ output. Do you have any idea?
• Serge Ballesta: @田中博貴: As I have already said in a comment, you are trying to read a text file as binary struct which is non sense...
• hiroyuki tanaka: @SergeBallesta: Thanks. It worked. C++ code and binary file are made by other person. maybe, I misunderstood about the difference between reading binary and text.