Re: Index key compression in informix?

----- Original Message ----- From: Rajib Sarkar <rsarkar@us.ibm.com> At: 2/12 12:45 This is OT, and you did say 'usually', but ... What you describe as the index key entry format is true for UNIQUE indexes only. Yes in a UNIQUE B+Tree index IDS stores Key, ROWID, Del flag for each key on the page/node, however, in a non-UNIQUE index the index format is an inversion list with the index entry being closer to key, ROWID list page address, empty key flag where the ROWID list page address points to the first of a chain of pages containing a ROWID and a Del flag for each row which has that key value. This inversion list format speeds Btree searches in non-unique indexes by eliminating duplicate node entries from the tree, especially in indexes with a high duplication rate. Just for information value. Art > The index format is usually Key, Rowid, Del flag ...we add the fragment > information too if its detached index on a fragmented table ... so if there > are duplicate keys, its not the key which is repeated but the ROWID ..:-) > > So, say, if the key is of datatype char(10) and the key is say "Test" then > within the index it will have the key as "Testbbbbbb" (where b is a blank > space) > > HTH > > > > Thanx much, > > Rajib Sarkar > Advisory Support Engineer (Wells Fargo Bank) > IBM Data Management Group > Ph : (602)-217-2100 > Fax: (602)-217-2100 > > As long as you derive inner help and comfort from anything, keep it -- > Mahatma Gandhi > > > > > "Bob Amy " > > <bamy@ssww.com> To: ids@iiug.org > > Sent by: cc: > > forum.subscriber@ Subject: Index key compression > in informix? [341] > iiug.org > > > > > > 02/12/2003 08:00 > > AM > > > > > > > > > > > Thanks for all the great info I see passing thru this mail list. You > guys are great! > > We have an indexed field where the last 20 characters are always blank > Does the index store all these blanks? > Can someone tell me if informix will truncate the trailing part of a key > that is always the same.