OpenCSV - Find out line number

Question

How do I know the actual line number on file associated to a line read by the CSVReader class? I can count the lines assuming every line read by this class is a new line on file. The problem is that it is possible to have newline characters in the CSV file. So for example, having 3 "logical" lines does not mean we have 3 "physical" lines on file. I have an error reporting feature that almost always report the wrong line number because of this.

Any ideas how to determine the real line number on file? Thanks!

Answer 1

If you are willing to modify the source code, you can add a counter to

private String getNextLine()

Increment the counter in the two places where

br.readLine();

is called, and expose the counter as a public property.

If you do not wish to modify the source code, for each CSV line returned, you could increment your own counter by 1 + the sum of the newline characters in the CSV line (presumably OpenCSV is returning columns including the newline characters to your code, though I have not tested that behavior). If you had Column A with one newline and Column B with two newlines, the actual file should look something like:

"This is

Cell A","And

Cell

B"

resulting in 3 newline characters (or \r\n sequences, depending on your platform), plus the 1 line returned by OpenCSV. Increment your counter by 4.

Answer 2

If you were willing to switch open source APIs to Super CSV (which distinguishes between the physical line and the CSV row), then you will have the following 3 methods available to you:

/**
 * Gets the current position in the file. 
 * The first line of the file is line number 1.
 * 
 * @since 1.0
 */
int getLineNumber();

/**
 * Returns the untokenized CSV row that was just read 
 * (which can potentially span multiple lines in the file).
 * 
 * @return the untokenized CSV row that was just read
 * @since 2.0.0
 */
String getUntokenizedRow();

/**
 * Gets the current row number (i.e. the number of CSV records - including the 
 * header - that have been read). This differs from the lineNumber, which is 
 * the number of real lines that have been read in the file. 
 * The first row is row 1 (which is typically the header row).
 * 
 * @since 2.0.0
 */
int getRowNumber();

Answer 3

You write that you need the line number for error reporting.
The CsvException class has a getLineNumber method that you can use.

Of course, this only works when you have an Exception.