Excel files that are different but equal

[ron]

I recently had a disk problem that chkdsk resolved by 'fixing' some things. When comparing to the master copy of the same data I found some differences in a few Excel files.

A BC 'data compare' says the files are the same. But a quick compare using CRC indicates they are different. Examining the hex view they are indeed different.

Presumably this portion of an Excel file is 'not important' or does not contain any of the actual 'data' from the spreadsheet. That's why BC's 'data compare' says the files are the same???

[Michael Bulgrien]

Even if you open an excel spreadsheet and, without making any changes, perform a "Save As" to another name, a binary compare will show that the new file on disk is different than the original. That is because office imbeds information such as the author, the file creation date, etc. in the file. As you suspected, that portion of the Excel file does not contain any of the actual 'data' from the spreadsheet.

For a data compare, BC3 extracts the textual data from the spreadsheets and, therefore, can verify that the content is the same even if there are differences in the non-data portion of the files.

[ron]

Ok, thanks. Still though, the files are different. I need to be more careful to use CRC when I'm comparing for backup purposes.

[Michael Bulgrien]

I usually opt for the "Binary comparison" option instead of the "CRC comparison". A CRC compare has to read both files in their entirety in order to calculate the CRC values for each file. The Binary compare compares files byte-by-byte and aborts the read on the first difference. For large files this can be much faster, especially if the the first difference is early in the file.

[ron]

Good point!

[Aaron]

This is very likely the case. I would suggest the Binary Comparison based on the above description.

Also, the BC3 Data Compare only compares 1 sheet at a time. You would want to use the Text Compare rule or setup a custom rule to compare the entire contents of multi-sheet files. See this KB article:

http://www.scootersoftware.com/support.php?zz=kb_multisheetexcel