Simple CSV Library

This is a CSV (delimiter can be changed) reader with a focus on: 1. Simplicity 2. Robustness 3. Performance (to a lesser extent)

It follows RFC 4180, but allows for non-conformant files to be processed. In order to accomplish this, it makes the following assumptions:

1. Commas on the end of a line results in a empty string for that column.
   • 1,2,3, is parsed as ["1","2","3",""]
2. Double quotes in a field that is not enclosed in double quotes are processed as a regular character and are included in the column string.
   • 1,2",3 is parsed as ["1","2\"","3"]
3. Non-delimiter characters immediately following a quoted field are treated as part of the column data and are appended to the column string.
   • 1,2,"3"123 is parsed as ["1","2","3123"]
4. An EOF in the middle of a quoted field is parsed as if the field was properly closed.
   • 1,2,"3*EOF* is parsed as ["1","2","3"]
5. There is no error for empty lines or varying number of columns per line.
   • An empty line is parsed as [""]

Limitations

• Lines are assumed to be UTF8 and are decoded "lossily" via Rust's String::from_utf8_lossy function.
• The iterator implementation forces an allocation for every row.
• The return character \r in unquoted fields is always discarded

Usage

Add to your Cargo.toml:

[dependencies] simple_csv = "~0.0.7"

Simple CSV usage

```rust let teststring = "1,2,3\r\n4,5,6".tostring(); let bytes = teststring.intobytes(); let mut testcsvreader = bytes.as_slice();

let mut reader = SimpleCsvReader::new(testcsvreader);

asserteq!(reader.nextrow(), Ok(vec!["1".tostring(),"2".tostring(),"3".tostring()].asslice())); asserteq!(reader.nextrow(), Ok(vec!["4".tostring(),"5".tostring(),"6".tostring()].asslice())); assert!(reader.nextrow().iserr()); ```

Different Delimiter

```rust let teststring = "1|2|3\r\n4|5|6".tostring(); let bytes = teststring.intobytes(); let testcsvreader = bytes.as_slice();

let mut reader = SimpleCsvReader::withdelimiter(testcsv_reader,'|');

asserteq!(reader.nextrow(), Ok(vec!["1".tostring(),"2".tostring(),"3".tostring()].asslice())); asserteq!(reader.nextrow(), Ok(vec!["4".tostring(),"5".tostring(),"6".tostring()].asslice())); assert!(reader.nextrow().iserr()); ```

Using a iterator

```rust let teststring = "1|2|3\r\n4|5|6".tostring(); let bytes = teststring.intobytes(); let testcsvreader = bytes.as_slice();

let mut reader = SimpleCsvReader::withdelimiter(testcsv_reader,'|');

for row in reader { println!("{}",row); } ```

Different Text Enclosing Character

```rust let teststring = "1,#2#,3\r\n#4#,5,6".tostring(); let bytes = teststring.intobytes(); let testcsvreader = bytes.as_slice();

let mut reader = SimpleCsvReader::withcustomchars(testcsvreader, ',', '#', '\n');

asserteq!(reader.nextrow(), Ok(vec!["1".tostring(),"2".tostring(),"3".tostring()].asslice())); asserteq!(reader.nextrow(), Ok(vec!["4".tostring(),"5".tostring(),"6".tostring()].asslice())); assert!(reader.nextrow().iserr()); ```

To Do

• Improve malformed UTF8 handling
• Implement CSV writer?
• Allow the iterator method to return errors somehow?