> -----Original Message----- > From: Bryan Pendleton [mailto:[EMAIL PROTECTED] > Sent: Tuesday, March 27, 2007 5:42 AM > To: Derby Discussion > Subject: Re: INPLACE Table Compression > > > "Inplace" Compression utility, no disk space is recovered and the size > > of the "*.dat" files is not reduced. > > That is correct. In-place compression re-arranges the records on > the existing pages of the existing file, gathering the existing > records together and shifting all the free space to be together. > This makes access to the existing records more efficient, and > allows new records to efficiently make use of the available free > space, but does nothing to reduce the overall size of the file. > > Think of it as somewhat similar to the "defragment" process that > your operating system offers. Your overall disk volume is still > the same size, but the use of the disk is improved. > > To release disk space back to the operating system, you must use > the version of compression which copies the data to a new file, > then deletes the old file.
Are you serious? Then surely the documentation is wrong. It says there are three operations: purge, defrag, and truncate; when truncate is used, it releases disk space to the operating system (by the way, it's not clearly stated, but I assume that defrag will do nothing unless a purge has been previously performed, and that truncate will do nothing unless a defrag has been previously performed). I'm looking at the 10.2.2 documentation (refderby.pdf, page 108). I haven't been able to test this for myself because in-place compress currently fails for me with ArrayIndexOutOfBoundsException. I assume this is due to DERBY-606 (though I don't have a table anywhere remotely near 2 GB), so I'm waiting to see if it works for us in 10.3. Jim
