Archives of the TeradataForum

Message Posted: Wed, 05 Nov 2003 @ 12:49:49 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Nancy,

My understanding is that when you convert a column from Latin to Unicode you basically double the internal server-side storage for that column. Latin uses 8-bit format, Unicode - 16-bit. If you convert from some Asian (Chinese, Japanese ...) character sets to Unicode, you don't need any extra storage as the old character sets were 16-bit anyway. This should help you estimate the costs for your columns.

On the client side, you have options ... If a client works with Latin characters only you can convert their data to Latin and send it as 8-bit. That would save some storage on the client side and improve performance. Again, no differences for 16-bit Asian clients storage- wise.

But my concern is more on the server side where storage is the most expensive. I have to deal with the languages which do have 8-bit encodings, such as Russian, Greek and Hebrew. It is difficult, however, to get them all together in _one_ encoding because the maximum number of characters you can have is only 256. That's why I am looking for some language specific workarounds.

Regards,

Victor

	<Prev	Next>		<<First	<Prev	Next>	Last>>