>>> about the fast export jdbc connectivity. is there any detailed instructions how to get the row count before fetch by not using select count(*)?
Select count(*) is the only way to obtain the row count prior to FastExport.
>>> When using native Fastexpoprt utility , this tool usually reports number of blocks being generated before the rows fetched from database .
There are three issues with the block count: (1) there is a variable number of rows in each block, so there is no way to estimate the row count from the block count; (2) the block count is only available with spooled FastExport, it's not available with a direct FastExport from a table that bypasses spool; and (3) the Teradata JDBC Driver does not provide the block count to a Java application.
>>> When using ODBC and .NET connections under SQLA, the resultset rowsize is also reported and visible before the rows were fetched. why not JDBC? it will be a good idea to fetch such info since it appears to me that it is available anyway.
SQLA displays the activity count, which is an approximate row count. There is feature request (JDBC RFC 121720) for the Teradata JDBC Driver to make the activity count available to an application. We haven't scheduled work on JDBC RFC 121720 yet.
↧