thanks guys, I found that the table is not partitioned, so I guess no way out...
On Mon, Sep 30, 2013 at 9:31 AM, Olga L. Natkovich <ol...@yahoo-inc.com>wrote: > You need to specify a table partition from which you want to sample.**** > > ** ** > > Olga**** > > ** ** > > *From:* Yang [mailto:teddyyyy...@gmail.com] > *Sent:* Sunday, September 29, 2013 1:39 PM > *To:* hive-u...@hadoop.apache.org > > *Subject:* how to treat an existing partition data file as a table?**** > > ** ** > > we have a huge table, including browsing data for the past 5 years, let's > say. **** > > ** ** > > now I want to take a few samples to play around with it. so I did**** > > select * from mytable limit 10;**** > > but it actually went full out and tried to scan the entire table. is there > a way to kind of create a "view" pointing to only one of the data files > used by the original table mytable ?**** > > this way the total files to be scanned is much smaller.**** > > ** ** > > ** ** > > thanks! > yang**** >
