Hi,

I'm profiling an application which processes large amounts of data and
have noticed that 25% of the time is spent doing the decode of bytes
into chars. I tried to find out what the actual code was behind
charsetdecoder but had no luck. However have other people found this
NIO class to be slow. When I say processing large amounts of data I
mean around 30gb's a day.

TIA

Niko wrote:
> Hi,
> 
> I'm profiling an application which processes large amounts of data and
> have noticed that 25% of the time is spent doing the decode of bytes
> into chars. I tried to find out what the actual code was behind
> charsetdecoder but had no luck. However have other people found this
> NIO class to be slow. When I say processing large amounts of data I
> mean around 30gb's a day.
> 
> TIA

It's converting bytes to unicode strings.

It decodes bytes using whatever characterset / character encoding scheme 
you have specified, or using the "platform default" if you have not 
specified. I owould guess that either ASCII or 8859_1 would be the 
fastest characterset to convert, if you're looking for speed-ups. But of 
course, if the text isn't really one of those two then you will run into 
characterset conversion problems.

Steve