Re: Basic help needed: how does indexing work...?!

From: Lars Grunewaldt <lg_at_netside.de>
Date: Wed, 13 Feb 2002 00:45:57 +0000
Message-ID: <3C69B745.7040706_at_netside.de>


Thx for replying.

Well, we just started the project, so we're not sure about anything, right now. Let's see:
- the databases are converted on PC side, so there is a fast processor for the index process
- the database is not planned to change on PDA. just uploaded and used. Maybe it'l come to "add this to your dictionary", but, not for now. If faster searches could be improved by dropping that, ... no problem. - the database records will be compressed, or, in fact, some may be, others not (depends on the database). Indexing could be done before or after, we use Huffman encoding, so the search string could be simply compressed by the same algorithm and be matched after compression - our database records might look something like this: varchar language1_1
varchar language1_2
varchar language2

maybe some integer/number values, but they're not so hard to compare, so I don't think we'll have to use an index there.

we'd like to search as comfortable as possible, so, really great would be an indexing mechanism which makes something like '%searchstring%' possible (or at least, like a wordmatch in fulltext search, '% searchstring %'). At least we'd implement 'searchstring%'. We're going to have large databases on small memory size, so the index size is important to. I just have no experiences in this sector, what makes it even difficult to define our database structures - I could change them "this" or "that", if I'd know what I'm looking for. So, thats why I ask for something like an "howto-Index" or so... but all I found on the web where things like "howto-develop-beginner-mysql&php-websites" - done that for 3 years now, so I know, basically, how indicies work from the user side. But from the database side....? *sigh*

hopes this helps to answer my questions :)

regards,

   lars

Robert Stearns wrote:

> I would need to know more about your needs before I could make a
> reasonable recommendation. Are you trying to index just the headwords of
> your dictionary or all of the definitions (quotes, examples, etc.; see
> OEDV2 on CD to see what complete indexing is like). You mentioned a PDA:
> are you going to compress your data, either before or after indexing? Is
> speed of retrieval more important either space or time of indexing? Is
> the dictionary: static; slow growth only; adds and deletes? Does the
> indexing have to be done on the PDA or will it be done on some type of
> faster processor?
> 
> Lars Grunewaldt wrote:
> 

>>OK, in short. I'm developing a database application (in fact, call it
>>dictionary). No, not just an application running on MySql, Sybase,
>>Ingres, Oracle. I *must* do this without (lets just say, PDA stuff. have
>>a look at sourceforge if you'd like. www.sourceforge.net/projects/padict).
>>
>>It may be possible that my database size gets so large that I'd need
>>indicies on my text entries. So, I would need some clues about "how to
>>index".
>>
>>I know a little bit about hash theories, but, what I need is a Book
>>advice, some newsgroups I could dig through, websites, whatever.
>>
>>I searched google, where I could find anything else 'til now.
>>
>>But: I'm stuck.
>>
>>If you know some good books, or websites, docs to read, samples,
>>opensource libraries...?
>>
>>regards,
>> Lars
>>
Received on Wed Feb 13 2002 - 01:45:57 CET

Original text of this message