Easiest way to compare two Excel files in Java?

Question

I\'m writing a JUnit test for some code that produces an Excel file (which is binary). I have another Excel file that contains my expected output. What\'s the easiest way to com

Answer 1

To test only content of the first sheets in Kotlin (easily can be converted to java).

private fun checkEqualityExcelDocs(doc : XSSFWorkbook, doc1 : XSSFWorkbook) : Boolean{
        val mapOfCellDoc = doc.toList().first().toList().flatMap { row -> row.map { Pair(PivotExcelCreator.IndexInThePivotTable(it.rowIndex,it.columnIndex),it.stringCellValue) }}.toMap()
        val mapOfCellDoc1 = doc1.toList().first().toList().flatMap { row -> row.map { Pair(PivotExcelCreator.IndexInThePivotTable(it.rowIndex,it.columnIndex),it.stringCellValue) }}.toMap()
        if(mapOfCellDoc.size == mapOfCellDoc1.size){
            return mapOfCellDoc.entries.all { mapOfCellDoc1.containsKey(it.key) && mapOfCellDoc[it.key] == mapOfCellDoc1[it.key]}
        }
        return false
    }

data class IndexInThePivotTable(val row: Int, val col: Int)

and in your code add assert

    assertTrue(checkEqualityExcelDocs(expected, actual), "Docs aren't equal!")

as you can see doc.toList().first() will take only the first sheet of document, if you need to compare each sheet respectively change code a little.

Also it is quite good idea to not take into account "" empty strings cells, I didn't need this functionality (As well, simply add this part, if you need).

---

also it can be useful information

//first doc I've got from outputstream such way
val out = ByteArrayOutputStream()
//some method which writes excel to outputstream
val firstDoc = XSSFWorkbook(ByteArrayInputStream(out.toByteArray()))

and second doc from file to compare with

val secondDoc = XSSFWorkbook(Test::class.java.getClassLoader().getResource("yourfile.xlsx").path)

Answer 2

Maybe... compare MD5 digests of each file? I'm sure there are a lot of ways to do it. You could just open both files and compare each byte.

EDIT: James stated how the XLS format might have differences in the metadata. Perhaps you should use the same interface you used to generate the xls files to open them and compare the values from cell to cell?

Answer 3

You might consider using my project simple-excel which provides a bunch of Hamcrest Matchers to do the job.

When you do something like the following,

assertThat(actual, WorkbookMatcher.sameWorkbook(expected));

You'd see, for example,

java.lang.AssertionError:
Expected: entire workbook to be equal
     but: cell at "C14" contained <"bananas"> expected <nothing>,
          cell at "C15" contained <"1,850,000 EUR"> expected <"1,850,000.00 EUR">,
          cell at "D16" contained <nothing> expected <"Tue Sep 04 06:30:00">
    at org.hamcrest.MatcherAssert.assertThat(MatcherAssert.java:20)

That way, you can run it from your automatted tests and get meaningful feedback whilst you're developing.

You can read more about it at this article on my site

Answer 4

I needed to do something similar and was already using the Apache POI library in my project to create Excel files. So I opted to use the included ExcelExtractor interface to export both workbooks as a string of text and asserted that the strings were equal. There are implementations for both HSSF for .xls as well as XSSF for .xlsx.

Dump to string:

XSSFWorkbook xssfWorkbookA = ...;
String workbookA = new XSSFExcelExtractor(xssfWorkbookA).getText();

ExcelExtractor has some options for what all should be included in the string dump. I found it to have useful defaults of including sheet names. In addition it includes the text contents of the cells.

Answer 5

Just found out there's something in commons-io's FileUtils. Thanks for the other answers.