Re: when to use a function or file for information

Agreed And if you had an index on all the fields, there would be a great disadvantage on your system. Think about this - what would the entry look like in the index? (What I'm going to describe is not exactly true as I understand it, but close enough for the concept.) It'd be a duplicate of each record - all the fields, with an extra field - the relative record number (or other pointer) to which the entry points. It is very conceivable that the index would be larger than the original table.

There is actually some optimization in the tree-page-node structure of an index, which is about all I'm really knowledgeable about, but I think the general idea holds. One optimization is that repeated data at the beginning of values is generally handled in one place, I think, so it's not repeated. That is, if you have values like "whiskey" and "whistle", one node would have "whis", then children with "key" and "tle". I think!! Unless I had too much whiskey and have started to whistle through my hat!

Vern

rob@xxxxxxxxx wrote:

True.

<snip>

The other advantage when selecting the data you need, is the query optimizer may use an IOA (Index Only Access), that means an index or logical file can found where all requested information is stored in the key fields, so no further access to the data base record is needed. IOA is the fastest method to access data. When using SELECT * an IOA will not be possible.

Do you mean if I have select clientnumber and an index on clientnumber the table will not be used, only the index?
Whereas if I had select *, I would need an index on all the fields for this to happen?
</snip>


Rob Berendt