Free open source CSV format parsing software written in C/C++
 

Google directed me to the ones written in .Net/C# etc.
Any ideas on the ones written in C/C++?

awk/gawk is witten in C. Just use: "awk -F ','".

Evo2.

---------- Post added 2011-06-01 at 17:39 ----------

awk/gawk is witten in C. Just use: "awk -F ','".

Evo2.

Thanks, that shows the power of awk! I found this: http://www.joeldare.com/wiki/using_awk_on_csv_files

Now I have a choice of calling awk through C by system system call. By that'll be a slow process. On the top of that I want the output in a string rather than a file.

I downloaded the awk source code and peeked in at the mammoth code :(

I've found this: http://code.google.com/p/csv-parser-cplusplus/. It looks promising.

and what have you been eating now-a-days for breakfast, BTW? It indeed looks promising, I've just untared it.

Just a couple of coffee cups! ;)

and I thought you must be having something special in the breakfast to be at your productive best! It is not easy to be on dot, always.

Ahh, sorry I didn't know you were looking for a library: thought you were looking for an executable.

Evo2.

Well, actually I was looking for a "small" program of 15-20 lines which I could embed in my code. :D

Anyway, you were helpful and thanks for that.

Which CSV formats do you want to parse? There's a good description of CSV formats here. Having delimiters and quotes in the data is common; having line ends in the data adds a level of complexity. If you are interested in deriving an algorithm for C++ from awk please ask and I'll post awk code (which does support line ends in the data).

Don't send C++ to do a camel's business ...

What about peeking at the 'awk' code in busybox? Or the code that handles IFS in some lightweight shell?

Umm, judging from wikipedia article it shouldn't be that hard to write parser from scratch in C++ (convert any input file into std::list<std::string> or std::vector<std::string> or std::vector<std::vector<std::string>>). using STL or Qt 4 the whole thing will probably take less than 100 (maybe even less than 50) lines. So, is there some kind of problem?

This one: http://www.robosuv.com/html/rddf_2004.html

Thanks, will go through that.

By "line ends" you mean some special characters denoting the end of line? Well that file doesn't have anything other than a '\n' char at the end.

and I am not sure that I have understood what you meant by "deriving an algorithm for C++ from awk". Did you mean that you'll be writing the code for the same in awk and then I'll translate the awk code to C++?
Anyway, I am an awk/sed/bash/perl illiterate. So won't be able to "read" anything you post in awk.

If you do bother to post the awk code [and also do write "what" each line is doing in plain English], it might be helpful in present/future, to either me or to someone else.

Thanks for your concern.

Not any more. Yesterday I went through some of the code of the software colucix posted. They appeared to be using some libraries, so I thought instead of spending time in exploring, understanding and then extracting code from their software, it is better to reinvent the wheel again. So, yesterday itself, I read the first row of the file [accidentally that was the only row in the file at that time :rolleyes:], put it in the std::string and extracted the substr from it on the basis of position returned by find function, of commas. Converted the string to double type by strtod function..

Didn't think that it would be too simple. Actually previously I had to parse an XML file in Qt and the Qt mailing list people told me to use a new class instead of regular one for parsing. And it really took quite an amount of time in writing that code, handling all the conditions etc.

XML is much more complicated than CSV.

Sounds like you forgot to handle quotes ("). A comma can be within quotes (","), and there can be quotes within quotes (""""). That's the only "difficulty", but it is still fairly trivial to implement. On other hand, if your file doesn't have any quotes, you don't need even that.