Skip blank lines while reading .csv file using opencsv (java)

Question

Good day everyone! My target is to make csv reader to skip the blank lines while parsing a file, do nothing basically, only get me the rows with at least one value. At the moment I have two methods -> 1st is just reading all rows as List of Strings array and returns it, 2nd converts the result into List of Lists of Strings, both are bellow:

private List<String[]> readCSVFile(File filename) throws IOException {

    CSVReader reader = new CSVReader(new FileReader(filename));
    List<String[]> allRows = reader.readAll();

    return allRows;

}

public List<List<String>> readFile(File filename) throws IOException {

        List<String[]> allRows = readCSVFile(filename);     
        List<List<String>> allRowsAsLists = new ArrayList<List<String>>();      
        for (String[] rowItemsArray :  allRows) {
            List<String> rowItems = new ArrayList<String>();
            rowItems.addAll(Arrays.asList(rowItemsArray));
            allRowsAsLists.add(rowItems);

        }
    return allRowsAsLists;

}

My first thought was to check (in the 2'nd method) the length of an array if its 0 just to ignore it - which would be something like this:

for (String[] rowItemsArray :  allRows) {
            **if(rowItemArray.length == 0) continue;**
            List<String> rowItems = new ArrayList<String>();
            rowItems.addAll(Arrays.asList(rowItemsArray));
            allRowsAsLists.add(rowItems);

}  

Unfortunately that didn't work for the reason that even if the row is blank it still returns an array of elements - empty Strings in fact. Checking an individual String is not an option as there are 100+ columns and this is variable. Please suggest what’s the best way to achieve this. Thanks.

Sorted it out this way:

    public List<List<String>> readFile(File filename) throws IOException {

            List<String[]> allRows = readCSVFile(filename, includeHeaders, trimWhitespacesInFieldValues);       
            List<List<String>> allRowsAsLists = new ArrayList<List<String>>();      
            for (String[] rowItemsArray :  allRows) {
                **if(allValuesInRowAreEmpty(rowItemsArray)) continue;**
                List<String> rowItems = new ArrayList<String>();
                rowItems.addAll(Arrays.asList(rowItemsArray));
                allRowsAsLists.add(rowItems);

            }
            return allRowsAsLists;

        }

    private boolean allValuesInRowAreEmpty(String[] row) {
        boolean returnValue = true;
        for (String s : row) {
            if (s.length() != 0) {
                returnValue = false;
            }
        }
        return returnValue;
    }

Answer 1

You could check the length and the first element. If the line contains only a field separator then the lenght > 1. If the line contains a single space character then the first element is not empty.

if (rowItemsArray.length == 1 && rowItemsArray[0].isEmpty()) {
    continue;
}

Answer 2

For opencsv 5.0 there is an API-option to read CSV lines directly into a Bean.

For people who prefer using the "CsvToBean" feature, the following solution is using the (sadly deprecated) #withFilter(..) method on CsvToBeanBuilder to skip blank lines in the Inputstream:

InputStream inputStream; // provided
List<MyBean> data = new CsvToBeanBuilder(new BufferedReader(new InputStreamReader(inputStream)))
    .withType(MyBean.class)
    .withFilter(new CsvToBeanFilter() {
        /*
         * This filter ignores empty lines from the input
         */
        @Override
        public boolean allowLine(String[] strings) {
            for (String one : strings) {
                if (one != null && one.length() > 0) {
                    return true;
                }
            }
            return false;
        }
    }).build().parse();

Update: With opencsv Release 5.1 (dated 2/2/2020), CsvToBeanFilter got undeprecated as per feature request #120.