A parser for csv files in C, for C and C++
This is an RFC 4180 parser for reading csv files for C and C++. It also handles csv files created by Excel, including multi line cells. The latest change allows it to read very large csv files ( tested to 500,000 rows )
The csv file is read into a linked list of rows, each row has a linked list of cells for that row. After creation of this list, an array of row pointers is created to allow fast lookup.
This allows each cell be access directly and in any order.
////////////////////////////////////// // The rows are a linked list // Each row has a linked list of cells // R - C - C // | // R - C -C - C - C - C // | // R - C -C - C - C // | // etc //////////////////////////////////////
There are just 3 functions, one to read the file and one to get access to the data in each cell, and one to free the memory used when done. It compiles fine using C or C++, so it can be used for C and C++.
CsvType *csvPtr = readCsv(filename, csvSeperator);
cell = getCell(csvPtr, row, column);
void freeMem(csvPtr);
See csvTest.c for a C example of reading and recreating the csv file to stdout.
For the C++ version, There is a simple class defined
See csvTest.cpp for a C++ example of reading and recreating the csv file to stdout.
I have written several csv parsers over the years. These parsers used the troublesome C functions 'strsep' and 'strtok' either I was using them wrong or they are buggy under a high load, they also behaved in different ways on different machines and OS's. This csv parser does the tokenising itself, which proved to be easier than getting 'strtok' to work reliably. The code reads the csv and into a tree like structure, a linked list of rows, each row has a linked list of cells. This makes it fast and easy to access each cell which can be accesses via row and column.
Whilst it is very fast, It should read files of any size, limited by how much memory you have. For extremely large files, it now uses an array of row pointers to quickly find the correct row. It has been tested on csv files with 500000 lines.
I used to roll my own linked lists before the C++ STL came along, and it was fun for me to write code using linked lists again.
I have tried to make it RFC 4180 compliant. This repository consists of two files csvParser.h csvParser.c
-- for testing and an example csvTest.c
The third file csvTest.c can be used as an example and for testing The csv file is read by this function
CsvType *csv = readCsv(filename, csvSeperator); Cells can be acceses via this function cell = getCell(csv, row, column);
for C++, there is a simple class defined, but it basically uses the c code. csvParser.h csvParser.cpp (This is the same file as csvParser.c) csvTest.cpp
The csv file is read by the this function CsvType *csv = readCsv(filename, csvSeperator); Cells can be acceses via this function cell = getCell(csv, row, column);
The returned cell data structure has enough information to tell is the cell exists, if it is empty and if it the last cell in the list (the end of the line) Anyway see csvParser.h and csvParser.c
See csvTest.c as an example
Memory Leaks? I could not measure any. :-)
Possible improvements:
- Allow single quotes as well as double quotes
- Windows based files will loose the the Line feed characters, this might be an issue for multiple line cells.