Does anyone have experience with using TRichView components with EDB?

If I store the contents of a TRichViewEdit control as RTF (Rich Text Format) or RVF (Rich View Format) in a CLOB field with a UNI collation, the text appears as what looks like Chinese or Japanese characters when viewed in EDB Manager.  Is this normal?

Also, when I do a text search of this field programmatically using SQL, it doesn't find the text I'm searching for (there is a text index for the field).  This is not too much of a problem because I also store the text as plain text in another field so I can search that field instead.

= Steve

Yes this is expected - TRichview has their own formatting codes embedded - so you have done the right thing if you want to do a text search by putting text into a seperate column in the db.  

If would be the same if you were to store a Word doc in the clob.

Walter

<< Yes this is expected - TRichview has their own formatting codes embedded - so you have done the right thing if you want to do a text search by putting text into a seperate column in the db.  

If would be the same if you were to store a Word doc in the clob. >>

Ok, thanks Walter.  It threw me a bit because when I stored RTF in a BLOB field in DBISAM you could actually read the text amongst the RTF codes using DBSys.

= Steve

Steve


As Walter said RichView uses its own custom format internally. I use WPTools and it allows you a choice of default IO format there may be something in RichView. Why not as the author  - I've forgotton his name - he was very helpful when I was looking at the alternatives.

As far as SQL searching goes with any sort of formatted wp document you're better off creating a full text index. This give the same advantage of removing the control codes that creating a separate column would do but give MUCH faster searching for words. Naturally it doesn't do so well for phrases but a combination of CONTAINS and LIKE or POS see below) should work.

Finally a point of warning. Tim has a wonderfully complete implementation of LIKE. It works extremely well for LIKE '%....' and '....%' but '%..%' on even medium sized CLOBS can be slow because of the sliding window needed to handle constructs such as '%this is a big?? search* than you would like%'. Since my needs are simple I switched to using POS which being simple like me is faster.


Roy Lambert

Steve Gill wrote:

> If I store the contents of a TRichViewEdit control as RTF (Rich Text
> Format) or RVF (Rich View Format) in a CLOB field with a UNI
> collation, the text appears as what looks like Chinese or Japanese
> characters when viewed in EDB Manager.  Is this normal?

To me, that sounds like it is Ansi text stored in a Unicode field.
That's why you are getting the unusual characters. RTF is all plain
text normally (Unicode characters are encoded as \U1341 or something
IIRC).

I'd either use a binary BLOB (not character blob) or make the text
Unicode string before storing.

I may be wrong though - I'm fairly new to ElevateDB.

--

Matthew Jones

>Steve Gill wrote:

I'd either use a binary BLOB (not character blob) or make the text
Unicode string before storing.
<

I would agree with Steve and store the TRichText in a Blob field in case it has embedded binary like an image or special formatting characters.

Barry

Hi Roy,

<< As far as SQL searching goes with any sort of formatted wp document you're better off creating a full text index. This give the same advantage of removing the control codes that creating a separate column would do but give MUCH faster searching for words. Naturally it doesn't do so well for phrases but a combination of CONTAINS and LIKE or POS see below) should work.  >>

Yes, I'm already using a text index and am using CONTAINS in the searches.

= Steve

Hi Matthew,

<  To me, that sounds like it is Ansi text stored in a Unicode field.
That's why you are getting the unusual characters. RTF is all plain
text normally (Unicode characters are encoded as \U1341 or something
IIRC).

I'd either use a binary BLOB (not character blob) or make the text
Unicode string before storing. >>

TRichViewEdit can save as either RVF or RTF, and in either binary or text format.  I have tried saving as both RVF and RTF in text format with the Unicode property set. I first wanted to find out if it was a limitation of EDB Manager before contacting the TRichView guys. I'll see what they have to say.

Thanks.

= Steve

Steve

><< As far as SQL searching goes with any sort of formatted wp document you're better off creating a full text index. This give the same advantage of removing the control codes that creating a separate column would do but give MUCH faster searching for words. Naturally it doesn't do so well for phrases but a combination of CONTAINS and LIKE or POS see below) should work. >>
>
>Yes, I'm already using a text index and am using CONTAINS in the searches.

Then is there a reason for "because I also store the text as plain text in another field"? Just curious

Roy Lambert

My suggestion applies only to a "plain text" RTF document, not to a
binary format of course. The binary wouldn't be searchable in EDB
either.

--

Matthew Jones