PC SOFT

FORUMS PROFESSIONNELS
WINDEVWEBDEV et WINDEV Mobile

Accueil → WINDEV 2024 → Full Text Index vs Database Design
Full Text Index vs Database Design
Débuté par SolutionJ-Reg, 11 juin 2017 12:09 - 7 réponses
Posté le 11 juin 2017 - 12:09
WB20

Hi, I'm keen to use the Full Text Index facility for a project I'm working on, but it looks like it can only be used on individual data files? I always design my data tables on 3rd normal form principles to avoid duplication of data storage, but it looks like I will have to compromise on this if I want to use the FTI facility, or am I missing something?

For example, I have an order table with foreign keys for elements like customer, delivery address etc, so the customer name is not available on the order table to be used in a Full Text Index.

Is there a way to use FTI's on a query for example?

Cheers
Reg
Posté le 11 juin 2017 - 22:49
Hi Reg,

I cannot answer your actual question, but would like to give you a heads up on Windev's FTI in general.

The full text index does not necessarily work even with single files.

I have a full text index on my contacts and products files. For some reason the product FTI seems to work OK, but the contact FTI finds hardly anything.

Best regards
Ola
Posté le 12 juin 2017 - 09:09
Hi,

It should work as you describe. Did you already try to create a query that way?
I'm using FTI's in some queries and it works pretty good.

Kind regards,

Joris.
Posté le 12 juin 2017 - 09:47
Hi,

I've decided to go about this in a different way and not use the FTI concept, but thanks for the input.

Kind Regards
Reg
Posté le 12 juin 2017 - 14:00
Hi

I tried and discarded the full text index provided by pcsoft because it's just too limited.

So I built my own full text indexing system, and it is available in my WXEDM open source project. Entirely written in wlanguage, you can adapt it and push it any way you want.

Best regards
Posté le 13 juin 2017 - 09:26
Hi Fabrice, I did the same for small files, the user can input up to three words to search for. It's fixed on "AND" and inserts the records found into a memory table. There are no additional fields / indices to define in the file. It sequentially reads all the of the records of the file, concatenates all the strings to search, does the search and inserts in case of "found". It's simple and of course, it works for small files only.
Posté le 13 juin 2017 - 14:02
Hi Guenter,

mine works with a dictionary file, supports exact and fuzzy searches, multilingual content, can index any field of any file, and works for ANY volume of data.

In WXEDM, I use it to fully index the documents content AND associated keywords, if any.

Best regards
Posté le 13 juin 2017 - 20:38
Hi Joris,

I thought that I can define a full text index for an analysis file only. How do you define the FTI for a query?

Best regards
Ola