File integrity checking?

[bogstick]

I like to routinely check my documents against backups by comparing MD5
signatures.
However, I find this impossible with Excel files because the MD5 value
changes just from opening and closing documents. I do not have this problem
with Word documents or any other type of file.
Could someone advise if there is an Excel setting that is causing this? If
so, do I need to use another file integrity checking method? What?

 

[bogstick]

Anyone???

What is causing Excel documents to give different MD5 checksum values when
nothing is changed other than opening and closing the file??

 

[Bob I]

If you do a FC (windows file compare) between a copy of the file that
has been opened and one that hasn't, what is listed as different. (.xlsx
files don't change for me)

 

[JP]

What exactly are you trying to do?

--JP

 

[bogstick]

You are right! This seems to be an issue only with Excel 2003 files!
When I used FC command, there was a small difference picked up!
I tried with 2007 .xlsx and the MD5 no longer changed!

So this is a fault of Excel 2003 documents? I wonder what the workaround
would be now if I were to still compare current Excel 2003 documents with
backup copies?

 

[bogstick]

I have some old backups of Excel 2003 documents and would like to verify
current copies against them! Therefore i am using the MD5 hashes for
comparing. But obviously this is failing because just opening and closing a
file causes the MD5 to change making the whole exercise useless! I do not
have any issues with any other type of Office 2003 documents.

 

[Bob I]

Ok, are you opening and closing the "Excel 2003" files with Excel 2007?
And if so, do they change every time you open and close them, or is it
just the first time.

 

[Bob I]

perhaps the modified date is not changed?

I have some old backups of Excel 2003 documents and would like to verify
current copies against them! Therefore i am using the MD5 hashes for
comparing. But obviously this is failing because just opening and closing a
file causes the MD5 to change making the whole exercise useless! I do not
have any issues with any other type of Office 2003 documents.

:

 

[JP]

If the MD5 hashes aren't matching up, I would check the file size and
last modified date instead.

--JP

 

[bogstick]

Yes i am opening Excel 2003 files with Excel 2007.
And yes the MD5 signature changes each time the file is opened and closed!
Are you able to replicate this with a blank document?

 

[bogstick]

No, the file size and modified dont change. I tried Excel 2003 files under
Excel 2003 and Excel 2007, and the issue affects both. How is it that FC is
picking up a difference yet the file size and last modified date remained
unchanged?

 

[JP]

I don't know, but I think you should stick with my suggestion.

--JP

 

[bogstick]

What suggestion?

I don't know, but I think you should stick with my suggestion.

--JP

 

[JP]

To check the file size and last modified date instead.

--JP

 

[Bob I]

With the FC command what changes are evident. Is this perhaps some
"accessed" time stamp?

 

[bogstick]

I already answered that!
See post 9/26/2009 12:16 Pm PST!

I said the file size and date modified do not change!
What this means is somehow the files have change just from being opened and
closed without affecting the above!

 

[bogstick]

The FC results are from what i remember about 9 or 10 hex characters?

I have just randomly opened and closed some Excel 2003 documents and the
Date Accessed do not change. Unless you're talking about some other Excel
specific timestamp? So are you not able to reproduce my issues?

 

[Bob I]

Using the /b switch it looks like 5 bytes change. the first at 046c
offset and the last at 470 offset. the first increments by 16 and the
last increments by one. I don't know the purpose. I think it is save to
say that if you are same modified date and file size then the two files
are the same worksheets.

 

[JP]

What I'm saying is that I don't understand your goal. Are you trying
to verify backup integrity, or compare Excel files using MD5 hashes?

If the MD5 hashes change constantly, use another method to compare
files, or continue to press the MD5 issue?

--JP

 

[Bogstick]

Yes i am trying to do a casual backup integrity check!
I use MD5 as its easy and quick to compare files between directories.
Excel 2003 is the only file format that fails.
In fact, i have tried a bunch of different hash signature checks and they
also fail!

Bob I above has already confirmed that there is a slight change (5 bytes)
just from opening and closing a file.

Are you suggesting there is something out there that i could use just for
this issue specifically??