Subject | Re: [ib-support] ato increment / IDENTITY |
---|---|
Author | Ann W. Harrison |
Post date | 2001-06-15T17:38:13Z |
At 07:00 PM 6/15/2001 +0200, Frank Ingermann wrote:
the hardware/firmware/software that handled network issues would
include compression. A high-level software compression seemed like
a lose. I really don't know - check my previous confession on the
subject of networks - what has happened in that direction.
Regards,
Ann
www.ibphoenix.com
We have answers.
>... But what about the network traffic?Somebody had to ask.
>- is the data stream from a query result sent across the wire in compressedFull length. At the time that decision was made, it appeared that
> form, and then decompressed by the gds32.dll on the client side, or is it
> decompressed by the server and then sent in "full length" to the client?
the hardware/firmware/software that handled network issues would
include compression. A high-level software compression seemed like
a lose. I really don't know - check my previous confession on the
subject of networks - what has happened in that direction.
>This would make a great difference for the choice of data typesNo, as it turns out, it doesn't.
>(Smallint vs. Numeric(18,0) included), but esp. between Char or VarChar.
>so how many bytes would go across the wire for: (assuming 1 byte per char)100 or 101 if the string is passed as a "c-string", null terminated.
>
>a) a CHAR(100) with 10 Chars (<> blank)
>b) a CHAR(100) with 100 Chars100 or 101 if the string is passed as a "c-string", null terminated.
>c) a VARCHAR(100) with 10 Chars102 or 101 if the string is passed as a "c-string", null terminated.
>d) a VARCHAR(100) with 100 Chars102 or 101 if the string is passed as a "c-string", null terminated.
Regards,
Ann
www.ibphoenix.com
We have answers.