|Informix Application Development - Getting Started|
|IBM Informix Developer's Handbook|
|Embedding IBM Informix|
Archive - May 2011
Update statistics low gathers information about an index by traversing
across the leaves of the btree. Prior to 11.70.xC3, Informix would
read each leaf page individually without the use of any readahead
functionality. This meant that for each leaf page encountered, unless
the leaf page happened to be in the bufferpool cache, the update
statistics low process would have to wait for that leaf page to be
brought into the bufferpool cache before processing it and moving to
the next leaf page.
When trying to find out the amount of raw data which is contain in the Informix instance there are two methods. The first is to determine the amount of data pages each table contains. This is reasonably accurate and very fast, it does have a draw back in that it must assume each page is full (or fix percentage of a data page is used). This method only has to examine the meta data for each table in the instance.