On Thursday, December 27, 2001, at 06:39 AM, Ruslan Zasukhin wrote:
> on 12/27/01 10:46, diskot123 /AT/ juno D.OT com at diskot123@juno.com wrote:
>
>>> Yes, VarChar size still play important role -- it regulate size of
>>> page for
>>> file. When you say 504 bytes, page is 1024 bytes. For VarChar[10000]
>>> page is
>>> 20KB. Since page is smallest atom to read / write from disk size of
>>> page will
>>> affect performance.
>> I've designed the database, and imported the data. The size of the .dat
>> file is fine even though I thought it would be high as I'm using
>> varchar[504] as you suggested. What is an issue however if the size of
>> the index file generated 4.8mb from 4 tables. How is the performance
>> with
>> LIKE as opposed to = which requires an index. Can valentina generate
>> the
>> index on the fly..and then delete it (automatically), to save the user
>> on
>> disk space?
>
> 1) LIKE in Valentina is again much faster than e.g. In mySQL, because
> field
> is logical file.
>
> 2) I can't say how much LIKE is slower of indexed search.
> Test self.
>
For my projects with V4RB, the time required for a RegEx search appears
to be dominated by the time it takes to put the results into a listbox,
so that the perceived difference between a RegEx search and an indexed
search is minimal, at most. I'm working with a 14 mb database
(excluding index).
Charles Yeomans
©2001 Charles Yeomans |
