CODEWITHSUNDEEP
javaexcelapache-poiexport-to-csv
Stone
Stone

Reputation: 673

Format Issue while reading excel data into Java

I am reading Excel data using java apache. I got format issue while reading double value such as 869.87929 (in excel) into 869.8792899999999 (in java).

I'm using following files to read excel data.
1. Schema.csv: SheetName,2-int-Double

2. File.xls:

col1 | col2 123 | 869.87929

Sample code: 

if(type.equals("Double")){
            Double fval=Double.parseDouble(content[i-1]);
            String sval=fval.toString();
            listObjects.add(new Double(Double.parseDouble(sval)));
        }

Note: type from schema.csv & content [] value from file.xls

Upvotes: 0

Views: 1441

Answers (3)

Peter Lawrey
Peter Lawrey

Reputation: 533432

There is no point converting number back and forth to Strings as this shouldn't do any thing useful.

Try doing

listObjects.add(new BigDecimal(content[i-1]));

with rounding you can do

listObjects.add(new BigDecimal(content[i-1]).setScale(9, RoundingMode.HALF_UP));

though I suspect the rounding error has occurred before this point as this should do basically the same thing as

listObjects.add(new Double(content[i-1]));

with rounding you can do

double d = Double.parseDouble(content[i-1]);
double round9 = Math.round(d * 1e9) / 1e9;
listObjects.add((Double) round9);

These are much the same as the number is within the precision of double and there should be no additional error here (i.e. the error is likely to be before this point)

Upvotes: 1

Kotodid
Kotodid

Reputation: 1179

If you use Apache POI - you can use getCellType()==Cell.CELL_TYPE_NUMERIC comparison and getNumericCellValue() from Cell interface.

Upvotes: 1

sashwat
sashwat

Reputation: 647

Double is not good for preserving precision. Preffered is using BigDecimal. I believe this is your problem.

https://blogs.oracle.com/CoreJavaTechTips/entry/the_need_for_bigdecimal

Upvotes: 1

Related Questions