my query had a bug in it ... it had on ON (it was repeated twice). try this
SELECT
h.sel_sid, h.hits, u.uniques, if(c.clicks is not null, c.clicks, 0)
FROM
selection_daily_hits h
left outer join selection_daily_uniques u ON (h.sel_sid = u.sel_sid AND
h.date_day = u.date_day AND u.date_day
Hello,
Here are the table descriptions. they only have the identifier, hits,
unqiques and date_day which is the partition
hive> describe selection_daily_hits;
OK
sel_sid int
hitsint
date_daystring
hive> describe selection_daily_uniques;
OK
sel_sid int
uniques int
date_daystri
I have a few questions as follows:
1) what's the schema of all 3 tables ? Do these tables only have (sel_sid,
date_day) as the columns along with the facts that they represent or do they
have more columns besides those 2 columns ?
2) why do you do a left outer join without an ON clause, not too
