Compression is a bad idea.
The file has to be uncompressed on the fly everytime you
access it.
Also, compression takes place in the background, and is
a pain in the butt, because you can't move or delete
files while the compression action is going on. Which
may take overnight as a background task.
Also, compression tends to make fragmentation even worse,
and defragmenting even more difficult.
Compact the MDB files, then zip them into zipped folders
(not into compressed folders), then delete the unzipped file.
If you have room, do the same to the SQL Databases. If not,
just zip and delete them. Do the zip/delete one file at
a time, so that the fragmentation does not just get worse.
A good third party defragmenter will defragment your
directory structure, which will make far more difference
than defragmenting your files ever will.
For these big files, you will get better speed if you
reformat with big clusters (but you won't be able to
defrag at all, so you shouldn't do this where you are
updating the data as the client is).
In fact, you will get better speed if you format your
40GB drive as FAT32, with very large clusters.
For the MDB files, connect using Exclusive. Also, try
connecting Read-only: I can't remember if that made
any difference.
And, consider re-indexing between searches. For some
kinds of searches, it is worth adding an index before
the search.
(david)
"Tom Ellison" <(E-Mail Removed)> wrote in message
news:%(E-Mail Removed)...
> Hey, Arvin,
>
> It's always a fine thing to have contact with you!
>
> There's no DVD drive on this system. The one on my other computer is
> temporarily unavailable as the HD there crashed completely. Have to
> rebuild a system AGAIN! Sigh!
>
> I think I'll compress the directories where these files are. Maybe then
> I'll be able to defrag one. If I get one, then maybe I'll be able to
> defrag another, and another.
>
> I had no idea the ability of Windows to defrag was so poor.
> Theoretically, a single file filling only 0.03 % of the hard drive, and
> being the only file on the hard drive, could be spread out across a drive
> an you'd be unable to defrag! The rule is, there must be a single
> contiguous free space large enough to hold the file or it won't defrag. A
> single 10 megabyte file on my 40 gigabyte drive could be too much for the
> defrag utility, if it were spread out in uniformly spaced chunks that are
> less than 10 megabytes apart. What a mess. I wrote my own defrag for unix
> systems about 20 years ago and it worked better than this one does!
> Really nuts!
>
> Tom Ellison
>
>
> "Arvin Meyer [MVP]" <(E-Mail Removed)> wrote in message
> news:%23HxOI$(E-Mail Removed)...
>> Hi Tom,
>>
>> My experience with MDBs is that you will get a 5 to 10% performance
>> increase
>> after the database is compacted. Compaction is not the same as disk
>> defragmentation, however, and I suspect that defragmenting will give you
>> a
>> similar improvement. With 19,000 fragments, you cannot help but improve
>> the
>> system condiserably. Indexing, as you already know, will give substantial
>> improvements in performance.
>>
>> You might suggest burning everything to a DVD if the client has one. I've
>> been buying 80GB USB drives for use in backup as cheap as $100 (and even
>> $50
>> dollars when there's a big sale).
>> --
>> Arvin Meyer, MCP, MVP
>> Microsoft Access
>> Free Access downloads
>> http://www.datastrat.com
>> http://www.mvps.org/access
>>
>> "Tom Ellison" <(E-Mail Removed)> wrote in message
>> news:(E-Mail Removed)...
>>> I have 6 very large and horribly fragmented database files (MDB and
>>> MDF).
>>> They range from 1.59 GB to 2.81 GB in size (SQL Express on the larger
>> ones).
>>> There are up to 19 thousand fragments in them. I've just learned that,
>> even
>>> though the HD is only 79% full, I cannot defrag these. I would have to
>>> somehow get 2.81 GB of CONTIGUOUS free space to do the largest. Looking
>> at
>>> the map in defrag, I doubt there is a single contiguous area of free
>>> space
>>> larger than a half GB.
>>>
>>> Before going to the expense and effort of buying another HD, installing
>> it,
>>> and copying these files, I want to ask your opinion and experience with
>> the
>>> performance issues involved. Is such a horribly fragmented MDB as this,
>>> scattered back and forth across a 40 GB drive, likely to see much
>>> performance improvement if defragged?
>>>
>>> I ran into a client who wanted performance. I decreased the storage of
>> most
>>> of the columns of data, reducing the size by 70% and more, then added an
>>> index. I got about a 10:1 performance improvement in queries. Now I
>>> want
>>> to ask the client to defrag what is a very large database (he's already
>>> reused most of the 70% savings by adding more records). I had hoped to
>>> defrag and test here, but that's being held up because I can't defrag.
>>>
>>> Do you see much performance improvement with a defrag on a database that
>> is
>>> in such bad condition? Any idea how much? The index still leaves a
>>> table
>>> scan of perhaps a million rows (rather than the total database of 10
>> million
>>> rows, hence the 10:1 improvement).
>>>
>>> I know, this is a crazy thing to do with Jet. I'm trying to get it
>> switched
>>> to SQL Express, but if I can illustrate an even greater improvement
>>> still
>>> using Jet, my credibility goes up. That will then tend to help move
>> things
>>> toward the ultimate solution.
>>>
>>> Tom Ellison
>>>
>>>
>>
>>
>
>