Hello.
Seeing the code around DERBY-326,
I guess that what makes the difference is that stream from clob is
re-encoded to UTF-8.
Original implementation expands stream from clob to memory entirely as
character,
and encode expanded characters to bytes array again.
I think saving the processes that expand large objects entirely to
memory twice improved the performance of streaming clob.
Best regards.
Oystein Grovlen - Sun Norway wrote:
TomohitoNakayama wrote:
Hello.
Yes for blob.
What makes blob different from clob? I would have guessed that the
implementation of clobs and blobs should be very similar.
--
/*
Tomohito Nakayama
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Naka
http://www5.ocn.ne.jp/~tomohito/TopPage.html
*/
