Reading binary files

Question

0.00/5 (No votes)

See more:

I currently have a VC6 app that takes a text file as input(records vary in length and have a unique field delimiter as well as CRLF). Below is the basic code:

string line;
ifstream infile("test.txt",ios::in);
while (!infile.eof())
{
   getline(infile,line);
}
...

The new input will have the same format(unique delimiters and LF retained), but be a binary file due to data requirements. If I change the above to open in binary mode, is this a valid means of reading a binary file line by line w/o data loss? The LF makes this possible even if it feels like it goes against the grain on how to read binary files.

Posted 26-Jan-11 12:28pm

zand29

Add a Solution

3 solutions

Solution 1

Technically, it should still work. But there will always be a danger of some new binary data being inserted that includes new line characters, which will mess up the file read.

Posted 26-Jan-11 12:38pm

Nish Nishant

Comments

Sergey Alexandrovich Kryukov 26-Jan-11 19:14pm

You see, you cannot give as answer without ambiguity like that. The final answer depends on what exactly is their idea behind the design of the representation of data in stream -- please see my answer. The issue is kind of "philosophical". (Oh, I'm afraid of aggressive reaction against my answer from some "readers", but... it's OK :-)
Thank you.
--SA

Solution 3

You should use Read/Write function for binary mode and ReadString/WriteString for text mode.

Posted 27-Jan-11 1:36am

Kozlov_Sergey

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Sergey Alexandrovich Kryukov · Accepted Answer · 2011-01-26T13:10:00

Here is the key: there is no such thing as non-binary data. "Text" data/file/mode is not an antonym of "binary", it's merely an indication that the data is processed in special way, which is system-dependent. So, this makes the statement binary file due to data requirements completely unclear.

In particular, making stream binary cannot harm unless you put end-of-line is some inconsistent way. But you're not going to! As I can see, you only want to complement your operation with binary stream output, nothing else.

The opposite is dangerous: it your "binary" format is designed to be system-independent and in strictly defined in the "every bit matters" stile, your text operation can mess up your system-independent "write line" operation. In this sense, it is not dangerous to use binary stream, it's dangerous to read/write it in system-dependent manner (and binary mode is designed to avoid it).

Another meaning of "text" is how numerical values are expressed -- this is obvious.

So, the final answer really depends on "data requirements" which you did not share with us. I would say, generally the idea of mixing up binary format with such text-specific options as "line" based on any kind of delimiters is extremely bad. This is a problem of data design, not implementation.

And this is much more important to clean up "data requirements" then choose a proper implementation. I would advice not even start implementation until possible issues of "data requirements" are made clear.

Based on my experience I would highly recommend to avoid using any delimiters in binary format. Every logical chunk of data should be prefixed in the stream by the length descriptor, so the implementation would know how many bytes to read from the stream at once into some structure. Usually the goal is to store/load of some data structure with a good performance because parsing is avoided — makes full sense.

Thank you.

—SA