#30 Using CSVBeanWriter to write empty Columns

[Anonymous]

Hi,

With CSVBeanReader the processiong of a column may omitted using "null" as headerColumn.
Doing the same thing with CSVBeanWriter leads to a NPE.

The Error occures at Line 67 of Version 1.6.0-SNAPSHOT.

I fixed it in my copy using following code instead of the lines 67 and 68:
if(methodName == null) {
result.add(null);
} else {
Method getMethod = cache.getGetMethod(source, methodName);
result.add(getMethod.invoke(source));
}

Using this and a ConverNullTo Processor enables writing empty/omitted columns.

[Anonymous]

Changed Version

[James Bassett]

Hi Ed,

I'll look into this, but my initial feeling is that it doesn't really make sense to use null as a header column. If you don't want to write the column, then why not remove it altogether from the nameMapping array? I don't believe you could actually write the header with writeHeader() if you have a null column name anyway.

James

[Anonymous]

Hi James,

in my case there is a defined column list, which must be supplied. So I have to add a column, which has no representation in my bean. Just like in the reader-case, where I can omit a column from the file and the bean doesn't have to have a dummy property.

I did the following:
- writeHeader with a String[] containing all header columns filled
- write with a String[], where the "omitted-to-be" columns were replaced with null

null within the writeHeader would lead to an NPE I guess.

regards
ed

[James Bassett]

I've updated CsvBeanWriter to allow for null entries in the nameMapping array ( see revision 231). This will result in an empty column (""). Also, note that I've updated all writers to write null as "" (previously you had to use a ConvertNullTo("") which was annoying).