On Fri, Feb 19, 2010 at 4:15 PM, Alan Winson <[email protected]> wrote:

> about 20% more unique keys than reading one file.  The lookup stage deletes
> each detail record that is added and adds a new detail record with the same
> key (and the updated sums).  I guess the reason that storage grows is either

I'll sit down with my pipeline wrench in a spare minute, but caution:
when you delete the master from lookup, the memory is not freed. That
has been on my wish list for a long time, but I doubt it will happen
this year for my birthday.

My paper on using lookup has an example of a pipeline that recycles
the lookup stage now and then to free memory and reload the valid
entries in the table.

Rob

Reply via email to