indexing: maxkeysize and pagesize

[Dominic Gamble <do...@stratlink.com.au>]

They maxkeysize and page size are parameters used when adding an index
to a collection.

Can anyone tell me exactly what these two parameters do, and what units
they expect?

I am assuming the larger the key size, the less chance of collisions
when putting the index entry in the hashtable, therefore better
performance when there are many keys. Is this correct?

is maxkeysize in bits? ie 32 bits, 64 bits??

What is pagesize and how does it affect indexing. Is it specified in
bytes?

Thanks for you help

Regards,
Dominic Gamble
Stratlink.

Re: indexing: maxkeysize and pagesize

[Jeff Greif <jg...@alumni.princeton.edu>]

I think (from a look at the source code) they pertain to how files
representing indexes are organized on the disk, and loaded into memory.  The
maxkeysize (in bytes, often a number like 256) specifies how long the key
assigned by the user or the system to a document in a collection is.  The
pagesize is usually a multiple of a file system blocksize convenient for
loading into memory or something similar.  Presumably the index is
represented as a b-tree or hash table with attribute or element text values
as the keys and document key strings as the values.

Jeff
----- Original Message -----
From: "Dominic Gamble" <do...@stratlink.com.au>
To: <xi...@xml.apache.org>
Sent: Wednesday, November 20, 2002 1:37 PM
Subject: indexing: maxkeysize and pagesize


> They maxkeysize and page size are parameters used when adding an index
> to a collection.
>
> Can anyone tell me exactly what these two parameters do, and what units
> they expect?
>
> I am assuming the larger the key size, the less chance of collisions
> when putting the index entry in the hashtable, therefore better
> performance when there are many keys. Is this correct?
>
> is maxkeysize in bits? ie 32 bits, 64 bits??
>
> What is pagesize and how does it affect indexing. Is it specified in
> bytes?
>
> Thanks for you help
>
> Regards,
> Dominic Gamble
> Stratlink.
>
>
>