RE: explanation of "out of space"

hen you delete rows, you don't release the disk space that was allocated. Inserting new rows will use existing holes in your table (provided you aren't using HPL in express mode or ploader in which case all loaded data is written to new disk and then those extents are appended to your table) - where was i. However the insertions will not use the 'holes' at 100% efficiency. To free up the space you 'created' when you deleted the rows you will need to rewrite the table. You will have to do this with an unload/drop/create/reload (given that you don't have the disk space to rewrite within the DB as done by altering an index to cluster). cheers j. > -----Original Message----- > From: M.J. Holley [mailto:datashark@fastlane.net] > Sent: Wednesday, November 29, 2000 6:54 AM > To: informix-list@iiug.org > Subject: explanation of "out of space" > > > Would someone please explain to me (or point me to the explanation) of > maxing out dbspace? About 2 months ago, I created a large > dbspace to add > room to a particular database (workgroup server, NT). > Yesterday I deleted > about 3 million records from a 5 million record table and updated > statistics. This morning, when trying to add a record, I am > told the record > can't be saved because I am out of disk space. It is NOT > physical hard > drive space. Theoretically I should have room for about 3 > million records in > that same table, but obviously it doesn't work this way. Some > clarification > would be appreciated. > > MJ > > > >

"Parker, Jack" <JParker@engage.com> writes: > hen you delete rows, you don't release the disk space that was allocated. > Inserting new rows will use existing holes in your table (provided you > aren't using HPL in express mode or ploader in which case all loaded data is > written to new disk and then those extents are appended to your table) - > where was i. However the insertions will not use the 'holes' at 100% > efficiency. To free up the space you 'created' when you deleted the rows > you will need to rewrite the table. You will have to do this with an > unload/drop/create/reload (given that you don't have the disk space to > rewrite within the DB as done by altering an index to cluster). An "ALTER FRAGMENT" command could be easier (not faster), and it's probably a good time to change those EXTENT SIZEs too. Art's myschema (with the -a switch) is a good help. Thomas