>
>
>> 1. Finding phantoms (AKA echos, I believe) -- words that have been
>> deleted from indexed VText fields but still are in the index.
>
>I hope this is fixed
No, I tried it yesterday (1.8.8r5) and it's not.
>
>
>> 2. Handling very large (> 30K) VText fields.
>
>What you mean?
>
>You can right now have 200-500KB text fields.
>Or you ask about Segment size ?
No, length. I'm sorry to say I've forgotten the exact details, but I
have a method that reads all of the fields in the record, converts
them to uppercase, and indexes them. That let's me do indexed,
case-insensitive searches for "words beginning with".
Each field (see below) is a VText field, so the method was declared
as such, too. The search was failing with large records, I believe,
and you told me to redelare it as VVarChar, which I did as
allFields = new VVarChar("allFields", 5000, -1)
allFields.indexByWords = true
allFields.SetMethod( "Upper(CONCAT(authors,' ', title,' ', editors,'
', journal,' ', volume,' ', pages,' ', date,' ', publisher,' ',
location,' ', url,' ',title2,' ',user1,' ',user2,' ', abstract,'
',keywords,' ',notes))" )
Of course, this isn't a long term solution, because only the 5000
characters will be available for searching.
>
>> 3. Appending one cursor to another (this isn't a bug like the others,
>> and I know that there's a workaround, but I have held off
>> implementing that in hopes that it will be added to Valentina in the
>> not-too-distant future).
>
>I keep this in mind. If we will lucky this will be in new SQL parser.
>If not, At least I will reserve this ability and later AFTER Server will
>return back to SQL
>
Fair enough. I guess I'll implement my own for now.
Thanks.
Jon
-------------------------------------------------------------
The Valentina mailing list is brought to you by MacServe.net
For info on lists services, see http://www.macserve.net/lists.html
©2001 jda |
|
